Search

This thesis studies the characteristics of data hosted on the interplanetary name system, which is a part of the interplanetary file system. From these records, information such as file names, locations, and sizes, was investigated. Data was collected on the number of peers hosting the records, thereby determining the decentralization of the record on the network. Data on how often content on the network changes, were collected and investigated. In addition to evaluating records, a search engine was prototyped to show how to integrate the data into a system. A large part of the network was crawled and the rate of change was found to be high. Most of the peers were found to host HTML files. Most content identifiers found were hosted by more than one peer. This means that a search engine needs to be able to support text file formats and revisit peers regularly to be up-to-date with the records.

A recently proposed defense for the anonymity network Tor uses preload lists of domains to determine what should be cached in the Domain Name System (DNS) caches of Tor relays. The defense protects against attacks that infer what is cached in Tor relays. By having domains continuously cached (preloaded), the cache will become independent of which websites have been visited. The current preload lists contain useless domains and have room for improvement. The objective of this project is to answer the question of "How can we generate better preload lists?" and to provide improved methods for generating preload lists, with the ultimate goal of generating better preload lists that the Tor Project can benefit from. We further developed existing tools to use web crawling to find more useful domains, as well as implementing filtering to remove useless domains from the preload lists. The results of this project showed promising results, as the useless domains decreased by an average of around 57% and more useful domains were found., Ett nyligen föreslaget försvar för anonymitetsnätverket Tor använder preload listor av domäner för att avgöra vad som ska cachelagras i domännamnssystemets (DNS) cacher för Tor reläer. Försvaret skyddar mot attacker som avgör vad som cachelagras i Tor relärer. Genom att ha domäner kontinuerligt cachade (förladdade), blir cachen oberoende av vilka websidor som har besökts. De nuvarande preload listorna innehåller värdelösa domäner och har utrymme för förbättring. Syftet med detta projekt är att svara på frågan "Hur kan vi generera bättre preload listor?" och att bidra med förbättrade metoder för att generera preload listor, med det ultimata målet att generera bättre preload listor som Torprojektet kan dra nytta av. Vi vidareutvecklade befintliga verktyg till att använda webbkrypning för att hitta mer användbara domäner, samt implementerade filtrering av värdelösa domäner från preload listorna. Detta projekt visade lovande resultat, då de värdelösa domänerna minskade med i genomsnitt ungefär 57% och mer användbara domäner hittades.

This thesis studies the characteristics of data hosted on the interplanetary name system, which is a part of the interplanetary file system. From these records, information such as file names, locations, and sizes, was investigated. Data was collected on the number of peers hosting the records, thereby determining the decentralization of the record on the network. Data on how often content on the network changes, were collected and investigated. In addition to evaluating records, a search engine was prototyped to show how to integrate the data into a system. A large part of the network was crawled and the rate of change was found to be high. Most of the peers were found to host HTML files. Most content identifiers found were hosted by more than one peer. This means that a search engine needs to be able to support text file formats and revisit peers regularly to be up-to-date with the records.

A recently proposed defense for the anonymity network Tor uses preload lists of domains to determine what should be cached in the Domain Name System (DNS) caches of Tor relays. The defense protects against attacks that infer what is cached in Tor relays. By having domains continuously cached (preloaded), the cache will become independent of which websites have been visited. The current preload lists contain useless domains and have room for improvement. The objective of this project is to answer the question of "How can we generate better preload lists?" and to provide improved methods for generating preload lists, with the ultimate goal of generating better preload lists that the Tor Project can benefit from. We further developed existing tools to use web crawling to find more useful domains, as well as implementing filtering to remove useless domains from the preload lists. The results of this project showed promising results, as the useless domains decreased by an average of around 57% and more useful domains were found., Ett nyligen föreslaget försvar för anonymitetsnätverket Tor använder preload listor av domäner för att avgöra vad som ska cachelagras i domännamnssystemets (DNS) cacher för Tor reläer. Försvaret skyddar mot attacker som avgör vad som cachelagras i Tor relärer. Genom att ha domäner kontinuerligt cachade (förladdade), blir cachen oberoende av vilka websidor som har besökts. De nuvarande preload listorna innehåller värdelösa domäner och har utrymme för förbättring. Syftet med detta projekt är att svara på frågan "Hur kan vi generera bättre preload listor?" och att bidra med förbättrade metoder för att generera preload listor, med det ultimata målet att generera bättre preload listor som Torprojektet kan dra nytta av. Vi vidareutvecklade befintliga verktyg till att använda webbkrypning för att hitta mer användbara domäner, samt implementerade filtrering av värdelösa domäner från preload listorna. Detta projekt visade lovande resultat, då de värdelösa domänerna minskade med i genomsnitt ungefär 57% och mer användbara domäner hittades.

Understanding the characteristics of tourists’ movements is essential for tourism destination management. With advances in informationand communication technology, more and more people are willing to upload photos and videos to various social media platforms while traveling. These openly available media data is gaining increasing attention in the field of movement pattern mining as a new data source. In this study, uploaded images and their geographic information within Lake Constance region, Germany were collected and through clustering analysis,a state-of-the-art k-means with noise removal algorithm was compared with the commonly used DBSCAN on Instagram dataset. Finally, association rules between popular attractions at region-level and citylevelwere mined respectively. Results show that social media data like Instagram constitute a valuable input to analyse tourists’ movement patterns as input to decision support and destination management.

© 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, This article introduces “Baby Robot”, a robot designed to improve infants’ and toddlers’ motor skills. This robot is a car-like toy that moves autonomously by using reinforcement learning and computer vision. Its behaviour consists of escaping from a target infant that has been previously recognized, or at least detected, without compromising the infant’s security by avoiding obstacles. Regarding other robots that share this purpose, there is a variety of commercial toys available on the market; however, no one is betting on an intelligent autonomous movement, since they use to repeat simple, yet repetitive movements. In order to examine how that autonomous movement may improve infants’ mobility, two crawling toys –one in representation of “Baby Robot”– were tested in a real environment. These real-life experiments were conducted with a safe and approved surrogate of our proposed robot in a kindergarten, where a group of infants interacted with the toys. Improvements in the efficiency of the play-session were detected., Peer Reviewed, Postprint (author's final draft)

Behavioral flexibility-the ability to tailor motor actions to changing body-environment relations-is critical for functional movement. Navigating the everyday environment requires the ability to generate a wide repertoire of actions, select the appropriate action for the current situation, and implement it quickly and accurately. We used a new, adjustable barrier paradigm to assess flexibility of motor actions in 20 17-month-old (eight girls, 12 boys) and 14 13-month-old (seven girls, eight boys) walking infants and a comparative sample of 14 adults (eight women, six men). Most participants were White, non-Hispanic, and middle class. Participants navigated under barriers normalized to their standing height (overhead, eye, chest, hip, and knee heights). Decreases in barrier height required lower postures for passage. Every participant altered their initial walking posture according to barrier height for every trial, and all but two 13-month-olds found solutions for passage. Compared to infants, adults displayed a wider variety of strategies (squat-walking, half-kneeling, etc.), found more appropriate solutions based on barrier height (ducked at eye height and low crawled at knee height), and implemented their solutions more quickly (within 4 s) and accurately (without bumping their heads against the barrier). Infants frequently crawled even when the barrier height did not warrant a low posture, displayed multiple postural shifts prior to passage and thus took longer to go, and often bumped their heads. Infants' improvements were related to age and walking experience. Thus, development of flexibility likely involves the contributions of multiple domains-motor, perception, and cognition-that facilitate strategy selection and implementation. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

This article introduces “Baby Robot”, a robot designed to improve infants’ and toddlers’ motor skills. This robot is a car-like toy that moves autonomously by using reinforcement learning and computer vision. Its behaviour consists of escaping from a target infant that has been previously recognized, or at least detected, without compromising the infant’s security by avoiding obstacles. Regarding other robots that share this purpose, there is a variety of commercial toys available on the market; however, no one is betting on an intelligent autonomous movement, since they use to repeat simple, yet repetitive movements. In order to examine how that autonomous movement may improve infants’ mobility, two crawling toys –one in representation of “Baby Robot”– were tested in a real environment. These real-life experiments were conducted with a safe and approved surrogate of our proposed robot in a kindergarten, where a group of infants interacted with the toys. Improvements in the efficiency of the play-session were detected., Peer Reviewed, Objectius de Desenvolupament Sostenible::5 - Igualtat de Gènere, Objectius de Desenvolupament Sostenible::10 - Reducció de les Desigualtats, Objectius de Desenvolupament Sostenible::3 - Salut i Benestar, Objectius de Desenvolupament Sostenible::4 - Educació de Qualitat, Postprint (published version)

Nowadays, online services, like e-commerce or streaming services, provide a personalized user experience through recommender systems. Recommender systems are built upon a vast amount of data about users/items acquired by the services. Such knowledge represents an invaluable resource. However, commonly, part of this knowledge is public and can be easily accessed via the Internet. Unfortunately, that same knowledge can be leveraged by competitors or malicious users. The literature offers a large number of works concerning attacks on recommender systems, but most of them assume that the attacker can easily access the full rating matrix. In practice, this is never the case. The only way to access the rating matrix is by gathering the ratings (e.g., reviews) by crawling the service's website. Crawling a website has a cost in terms of time and resources. What is more, the targeted website can employ defensive measures to detect automatic scraping. In this paper, we assess the impact of a series of attacks on recommender systems. Our analysis aims to set up the most realistic scenarios considering both the possibilities and the potential attacker's limitations. In particular, we assess the impact of different crawling approaches when attacking a recommendation service. From the collected information, we mount various profile injection attacks. We measure the value of the collected knowledge through the identification of the most similar user/item. Our empirical results show that while crawling can indeed bring knowledge to the attacker (up to 65% of neighborhood reconstruction on a mid-size dataset and up to 90% on a small-size dataset), this will not be enough to mount a successful shilling attack in practice., Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Cyber Security

Understanding the characteristics of tourists’ movements is essential for tourism destination management. With advances in informationand communication technology, more and more people are willing to upload photos and videos to various social media platforms while traveling. These openly available media data is gaining increasing attention in the field of movement pattern mining as a new data source. In this study, uploaded images and their geographic information within Lake Constance region, Germany were collected and through clustering analysis,a state-of-the-art k-means with noise removal algorithm was compared with the commonly used DBSCAN on Instagram dataset. Finally, association rules between popular attractions at region-level and citylevelwere mined respectively. Results show that social media data like Instagram constitute a valuable input to analyse tourists’ movement patterns as input to decision support and destination management.

In this paper, a multi-functional soft robot module that can be used to constitute a variety of soft robots is proposed. The body of the soft robot module made of rubber is in the shape of a long strip, with cylindrical chambers at both the top end and bottom end of the module for the function of actuators and sensors. The soft robot module is driven by supercoiled polymer artificial muscle (SCPAM) strings, which are made from conductive nylon sewing threads. Artificial muscle strings are embedded in the chambers of the module to control its deformation. In addition, SCPAM strings are also used for the robot module’s sensing based on the linear relationship between the string’s length and their resistance. The bending deformation of the robot is measured by the continuous change of the sensor’s resistance during the deformation of the module. Prototypes of an inchworm-like crawling robot and a soft robotic gripper are made, whose crawling ability and grasping ability are tested, respectively. We envision that the proposed proprioceptive soft robot module could potentially be used in other robotic applications, such as continuum robotic arm or underwater robot. © 2022 by the author. Licensee MDPI, Basel, Switzerland.

Learning programming boils down to the practice of solving exercises. However, although there are good and diversified exercises, these are held in proprietary systems hindering their interoperability. This article presents a simple scraping tool, called ScraPE, which through a navigation, interaction and data extraction script, materialized in a domain-specific language, allows extracting the data necessary from Web pages - typically online judges - to compose programming exercises in a standard language. The tool is validated by extracting exercises from a specific online judge. This tool is part of a larger project where the main objective is to provide programming exercises through a simple GraphQL API.

Understanding the characteristics of tourists’ movements is essential for tourism destination management. With advances in informationand communication technology, more and more people are willing to upload photos and videos to various social media platforms while traveling. These openly available media data is gaining increasing attention in the field of movement pattern mining as a new data source. In this study, uploaded images and their geographic information within Lake Constance region, Germany were collected and through clustering analysis,a state-of-the-art k-means with noise removal algorithm was compared with the commonly used DBSCAN on Instagram dataset. Finally, association rules between popular attractions at region-level and citylevelwere mined respectively. Results show that social media data like Instagram constitute a valuable input to analyse tourists’ movement patterns as input to decision support and destination management.

Understanding the characteristics of tourists’ movements is essential for tourism destination management. With advances in informationand communication technology, more and more people are willing to upload photos and videos to various social media platforms while traveling. These openly available media data is gaining increasing attention in the field of movement pattern mining as a new data source. In this study, uploaded images and their geographic information within Lake Constance region, Germany were collected and through clustering analysis,a state-of-the-art k-means with noise removal algorithm was compared with the commonly used DBSCAN on Instagram dataset. Finally, association rules between popular attractions at region-level and citylevelwere mined respectively. Results show that social media data like Instagram constitute a valuable input to analyse tourists’ movement patterns as input to decision support and destination management.

Understanding the characteristics of tourists’ movements is essential for tourism destination management. With advances in informationand communication technology, more and more people are willing to upload photos and videos to various social media platforms while traveling. These openly available media data is gaining increasing attention in the field of movement pattern mining as a new data source. In this study, uploaded images and their geographic information within Lake Constance region, Germany were collected and through clustering analysis,a state-of-the-art k-means with noise removal algorithm was compared with the commonly used DBSCAN on Instagram dataset. Finally, association rules between popular attractions at region-level and citylevelwere mined respectively. Results show that social media data like Instagram constitute a valuable input to analyse tourists’ movement patterns as input to decision support and destination management.

Users of anonymity networks face differential treatment and sometimes get blocked by websites, it is currently unclear how common this blocking is. This research aims to provide an overview of how common this blocking is while utilizing the AN.ON anonymity network. The analysis is accomplished by utilizing automated web scraping and processing to recognise and classify blocks by comparing them to a control connection. This process and software can be used and extended to analyze and compare any two connections. The scope is limited to the one thousand most popular websites according to the Alexa rating. Different kinds of blocks were identified and automatically recognised in processing, though manual verification is still required. Evidence is found and presented that there is a significant amount of blocking, occurring on approximately 23% of the analyzed domains. There is also a significant difference in blocking between using different cascades., CSE3000 Research Project, Computer Science and Engineering

© 2020 John Wiley & Sons Ltd Infants’ avoidance of drop-offs has been described as an affordance learning that is not transferable between different locomotor postures. In addition, there is evidence that infants perceive and act similarly around real and water cliffs. This cross-sectional study investigated the effects of specific locomotor experiences on infants’ avoidance behaviour using the Real Cliff/Water Cliff paradigm. The experiments included 102 infants, 58 crawling, but pre-walking, infants (Mage = 11.57 months, SD = 1.65) with crawling experience ranging between 0.03 and 7.4 months (M = 2.16, SD = 1.71) and 44 walking infants (Mage = 14.82 months, SD = 1.99), with walking experience ranging between 0.13 and 5.2 months (M = 1.86, SD = 1.28). The association between crawling experience and crawlers’ avoidance of the real and water cliffs was confirmed. Importantly, crawling and total self-produced locomotor experience, and not walking experience, were associated with walkers’ avoidance behaviour on both cliffs. These results suggest that some degree of perceptual learning acquired through crawling experience was developmentally transferred to the walking posture. A longer duration of crawling experience facilitates a more rapid recalibration to the new walking capability. In addition, there was no difference in infants’ avoidance of falling on the real and the water cliff. However, infants explored the water cliff more than the real cliff, revealing more enticement to examine bodies of water than for drop-offs.

Users of anonymity networks face differential treatment and sometimes get blocked by websites, it is currently unclear how common this blocking is. This research aims to provide an overview of how common this blocking is while utilizing the AN.ON anonymity network. The analysis is accomplished by utilizing automated web scraping and processing to recognise and classify blocks by comparing them to a control connection. This process and software can be used and extended to analyze and compare any two connections. The scope is limited to the one thousand most popular websites according to the Alexa rating. Different kinds of blocks were identified and automatically recognised in processing, though manual verification is still required. Evidence is found and presented that there is a significant amount of blocking, occurring on approximately 23% of the analyzed domains. There is also a significant difference in blocking between using different cascades., CSE3000 Research Project, Computer Science and Engineering

Актуальність теми дослідження. Сучасний широкий інтернет є істотним джерелом даних для використання у наукових та бізнес-дослідженнях. Можливість видобувати актуальні дані часто є ключовою для досягнення необхідних цілей, але сучасні якісні рішення з застосуванням технологій машинного зору та інших можуть бути дорогими до придбання або розробки, тому прості та дешеві як з точки зору розробки та підтримки, так і з точки зору експлуатації рішення є необхідними. Метою дослідження є створення програмного інструментарію екстракції структурованих даних з веб-сторінок новинних ресурсів для подальшої класифікації за достовірністю. Для досягнення поставленої мети було окреслено та виконано наступні завдання: - провести огляд існуючих підходів та програмних аналогів у областях екстракції даних з веб-ресурсів та оцінки якості новин; - позробити та реалізувати алгоритми екстракції, підготовки та класифікації даних; - порівняти результати, отримані розробленим алгоритмом та результатами тренування алгоритмів машинного навчання на даних, видобутих ним з існуючим аналогом та результатами тренування на даних аналогу. Об’єктом дослідження є процес екстракції текстових даних з подальшою обробкою методами машинного навчання. Предметом дослідження є методи та засоби екстракції та аналізу структурованих текстових даних. Наукова новизна одержаних результатів. Було створено простий жадібний алгоритм у якому суміщено процеси пошуку посилань та видобування інформації, доведено доцільність використання простих алгоритмів для збору даних з ресурсів у мережі Інтернет з ціллю використання у тренуванні алгоритмів машинного навчання. Було доведено що як класичні алгоритми навчання здатні досягати результатів, співставним з такими у нейронних мереж, таких як мережі ДКЧП, та показано що такі моделі здатні працювати на двомовному датасеті. Публікації. Матеріали роботи було опубліковано у п’ятій Всеукраїнській науково-практичній конференції молодих вчених та студентів «Інформаційні системи та техноло, Relevance of the research topic. Modern wide internet is a considerable source of data to be used in scientific and business applications. An ability to extract up to date data is frequently crutial for reaching necessary goals, though, modern quality solutions to this problem, which are using computer vision and other technologies, may be finantially demanding to acquire or develop, thus simple and cheap to develop, maintain and use solutions are necessary. The purpose of the study is to create a software instrument aimed at extraction of structured data from news websites for usage in news trustworthiness classification. Following tasks were outlined and implemented to achieve the aforementioned goal: - Outline existing approaches and analogues in areas of data extraction and news classification; - Design and develop extraction, preparation and classification algorhitms; - Compare the results achieved with developed extraction algorhitm and with existing software solution, including comparing machine learning accuracies on both of the extractors. The object of the study is the process of text data extraction with subsequent machine learning analysis. The subjects of the study are methods and tools of extraction and analysis of text data. Scientific novelty of the obtained results. A simple greedy algorithm was created, combining the process of link discovery and data extraction. Expediency of usage of simple web data extraction algorithms for composing machine learning datasets was proven. It was also proven that classical machine learning algorithms can achieve results similar to neural networks such as LSTM. Capabilities of machine learning systems to function efficiently in a bilingual context were also shown. Publications. Materials, related to this study, were published in the All-Ukrainian Scientific and Practical Conference of Young Scientists and Students “Information Systems and Management Technologies” (ISTU-2019) “News trustworthiness classification wit

© 2020 John Wiley & Sons Ltd Infants’ avoidance of drop-offs has been described as an affordance learning that is not transferable between different locomotor postures. In addition, there is evidence that infants perceive and act similarly around real and water cliffs. This cross-sectional study investigated the effects of specific locomotor experiences on infants’ avoidance behaviour using the Real Cliff/Water Cliff paradigm. The experiments included 102 infants, 58 crawling, but pre-walking, infants (Mage = 11.57 months, SD = 1.65) with crawling experience ranging between 0.03 and 7.4 months (M = 2.16, SD = 1.71) and 44 walking infants (Mage = 14.82 months, SD = 1.99), with walking experience ranging between 0.13 and 5.2 months (M = 1.86, SD = 1.28). The association between crawling experience and crawlers’ avoidance of the real and water cliffs was confirmed. Importantly, crawling and total self-produced locomotor experience, and not walking experience, were associated with walkers’ avoidance behaviour on both cliffs. These results suggest that some degree of perceptual learning acquired through crawling experience was developmentally transferred to the walking posture. A longer duration of crawling experience facilitates a more rapid recalibration to the new walking capability. In addition, there was no difference in infants’ avoidance of falling on the real and the water cliff. However, infants explored the water cliff more than the real cliff, revealing more enticement to examine bodies of water than for drop-offs.

Over the past 15 years, YouTube has emerged as a large and dominant social media service, giving rise to a ‘platformed media system’ within its technical and regulatory infrastructures. This paper relies on a large-scale sample of channels (n=36M+) to explore this media system along three main lines. First, we investigate stratification and hierarchization in broadly quantitative terms, connecting to well-known tropes on structural hierarchies emerging in networked systems, where a small number of elite actors often dominate visibility. Second, we inquire into YouTube’s channel categories, their relationships, and their proportions as a means to better understand the topics on offer and their relative importance. Third, we analyze channels according to country affiliation to gain insights into the dynamics and fault lines that align with country and language. Throughout the paper, we emphasize the inductive character of this research, by highlighting the many follow-up questions that emerge from our findings.

[Abstract] A chatbot is a type of agent that allows people to interact with an information repository using natural language. Nowadays, chatbots have been incorporated in the form of conversational assistants on the most important mobile and desktop platforms. In this article, we present our design of an assistant developed with open-source and widely used components. Our proposal covers the process end-to-end, from information gathering and processing to visual and speech-based interaction. We have deployed a proof of concept over the website of our Computer Science Faculty.

[ES] La generación automática de resúmenes es un campo muy atractivo y vigente dentro del área del procesamiento del lenguaje natural. Por esta razón, en este trabajo, se han decidido analizar y comparar distintas técnicas para la generación automática de resúmenes tanto basadas en redes neuronales como basadas en algoritmos clásicos. Para la evaluación de estas técnicas, se utilizan métricas sintácticas usadas en la mayoría de estudios de este campo y métricas semánticas propuestas en este mismo trabajo. Mediante este estudio, se podrán observar las diferencias que existen entre las dos evaluaciones comentadas anteriormente y la mejora que aporta la medida semántica propuesta. Para la utilización de aproximaciones de resumen basadas en redes neuronales y para la evaluación de todos los sistemas, se requiere un corpus de documentos y resúmenes. Por este motivo, mediante un proceso de crawling, se ha elaborado un corpus de noticias que provienen de distintos sitios web de prensa digital y para los idiomas de castellano y catalán. Finalmente, el trabajo contiene un extenso análisis de los resultados experimentales obtenidos en los corpus. En este, se podrán observar las diferencias que existen entre las noticias que contienen ambos corpus utilizando para ello distintas medidas. Además, se compararán los corpus generados con otros corpus que forman parte del estado del arte., [CA] La generació automàtica de resums és un camp molt atractiu i vigent dins de l’àrea del processament del llenguatge natural. Per aquesta raó, en aquest treball, s’ha decidit analitzar i comparar distintes tècniques per a la generació automàtica de resums, tant basades en xarxes neuronals com basades en algoritmes clàssics. Per a l’avaluació d’aquestes tècniques es fan servir mètriques sintàctiques utilitzades en la major part d’estudis d’aquest camp i mètriques semàntiques propostes en aquest treball. Mitjançant aquest estudi, es podran observar les diferències que existeixen entre les dues avaluacions comentades anteriorment i la millora que aporta la mesura semàntica proposta. Per a l’ús d’aproximacions de resum basades en xarxes neuronals i per a l’avaluació de tots els sistemes, es requereix un corpus de documents i resums. Per aquest motiu, mitjançant un procés de crawling, s’ha elaborat un corpus de notícies que provenen de distints llocs web de premsa digital i per als idiomes de castellà i català. Finalment, el treball conté un extens anàlisi dels resultats experimentals obtinguts en el corpus. En aquest es podran observar les diferències que existeixen entre les notícies que contenen ambdós corpus emprant distintes mesures. A més, es compararan els corpus generats amb altres corpus que formen part del estat de l’art., [EN] Automatic summary generation is a current and very attractive field in the area of natural language processing. For this reason, in this work, it has been decided to analyse and compare different techniques for automatic abstract generation both based on neural net-work and based on classical algorithms. For the evaluation of these techniques, syntactic metrics used in most studies of this field and semantic metrics proposed in this work have been used. Through this study, differences between the two evaluations discussed above can be seen. Moreover, improvements introduces by the semantic measure proposal are also presented. For the use of summary approaches based on neural networks and for the evaluation of all systems, a corpus of documents and summaries is required. For this reason, through a crawling process, a corpus of news that come from different digital press websites has been prepared in both Spanish and Catalan languages. Finally, the work contains an extensive analysis of the experimental results obtained with the corpus. In it, the existing differences between the news contained in both corpus can be observed. At the same time, a comparison of the different measures has been included. In addition, the generated corpus will be compared with other corpus that take part of the state of the art.

[ES] La generación automática de resúmenes es un campo muy atractivo y vigente dentro del área del procesamiento del lenguaje natural. Por esta razón, en este trabajo, se han decidido analizar y comparar distintas técnicas para la generación automática de resúmenes tanto basadas en redes neuronales como basadas en algoritmos clásicos. Para la evaluación de estas técnicas, se utilizan métricas sintácticas usadas en la mayoría de estudios de este campo y métricas semánticas propuestas en este mismo trabajo. Mediante este estudio, se podrán observar las diferencias que existen entre las dos evaluaciones comentadas anteriormente y la mejora que aporta la medida semántica propuesta. Para la utilización de aproximaciones de resumen basadas en redes neuronales y para la evaluación de todos los sistemas, se requiere un corpus de documentos y resúmenes. Por este motivo, mediante un proceso de crawling, se ha elaborado un corpus de noticias que provienen de distintos sitios web de prensa digital y para los idiomas de castellano y catalán. Finalmente, el trabajo contiene un extenso análisis de los resultados experimentales obtenidos en los corpus. En este, se podrán observar las diferencias que existen entre las noticias que contienen ambos corpus utilizando para ello distintas medidas. Además, se compararán los corpus generados con otros corpus que forman parte del estado del arte., [CA] La generació automàtica de resums és un camp molt atractiu i vigent dins de l’àrea del processament del llenguatge natural. Per aquesta raó, en aquest treball, s’ha decidit analitzar i comparar distintes tècniques per a la generació automàtica de resums, tant basades en xarxes neuronals com basades en algoritmes clàssics. Per a l’avaluació d’aquestes tècniques es fan servir mètriques sintàctiques utilitzades en la major part d’estudis d’aquest camp i mètriques semàntiques propostes en aquest treball. Mitjançant aquest estudi, es podran observar les diferències que existeixen entre les dues avaluacions comentades anteriorment i la millora que aporta la mesura semàntica proposta. Per a l’ús d’aproximacions de resum basades en xarxes neuronals i per a l’avaluació de tots els sistemes, es requereix un corpus de documents i resums. Per aquest motiu, mitjançant un procés de crawling, s’ha elaborat un corpus de notícies que provenen de distints llocs web de premsa digital i per als idiomes de castellà i català. Finalment, el treball conté un extens anàlisi dels resultats experimentals obtinguts en el corpus. En aquest es podran observar les diferències que existeixen entre les notícies que contenen ambdós corpus emprant distintes mesures. A més, es compararan els corpus generats amb altres corpus que formen part del estat de l’art., [EN] Automatic summary generation is a current and very attractive field in the area of natural language processing. For this reason, in this work, it has been decided to analyse and compare different techniques for automatic abstract generation both based on neural net-work and based on classical algorithms. For the evaluation of these techniques, syntactic metrics used in most studies of this field and semantic metrics proposed in this work have been used. Through this study, differences between the two evaluations discussed above can be seen. Moreover, improvements introduces by the semantic measure proposal are also presented. For the use of summary approaches based on neural networks and for the evaluation of all systems, a corpus of documents and summaries is required. For this reason, through a crawling process, a corpus of news that come from different digital press websites has been prepared in both Spanish and Catalan languages. Finally, the work contains an extensive analysis of the experimental results obtained with the corpus. In it, the existing differences between the news contained in both corpus can be observed. At the same time, a comparison of the different measures has been included. In addition, the generated corpus will be compared with other corpus that take part of the state of the art.

[ES] La generación automática de resúmenes es un campo muy atractivo y vigente dentro del área del procesamiento del lenguaje natural. Por esta razón, en este trabajo, se han decidido analizar y comparar distintas técnicas para la generación automática de resúmenes tanto basadas en redes neuronales como basadas en algoritmos clásicos. Para la evaluación de estas técnicas, se utilizan métricas sintácticas usadas en la mayoría de estudios de este campo y métricas semánticas propuestas en este mismo trabajo. Mediante este estudio, se podrán observar las diferencias que existen entre las dos evaluaciones comentadas anteriormente y la mejora que aporta la medida semántica propuesta. Para la utilización de aproximaciones de resumen basadas en redes neuronales y para la evaluación de todos los sistemas, se requiere un corpus de documentos y resúmenes. Por este motivo, mediante un proceso de crawling, se ha elaborado un corpus de noticias que provienen de distintos sitios web de prensa digital y para los idiomas de castellano y catalán. Finalmente, el trabajo contiene un extenso análisis de los resultados experimentales obtenidos en los corpus. En este, se podrán observar las diferencias que existen entre las noticias que contienen ambos corpus utilizando para ello distintas medidas. Además, se compararán los corpus generados con otros corpus que forman parte del estado del arte., [CA] La generació automàtica de resums és un camp molt atractiu i vigent dins de l’àrea del processament del llenguatge natural. Per aquesta raó, en aquest treball, s’ha decidit analitzar i comparar distintes tècniques per a la generació automàtica de resums, tant basades en xarxes neuronals com basades en algoritmes clàssics. Per a l’avaluació d’aquestes tècniques es fan servir mètriques sintàctiques utilitzades en la major part d’estudis d’aquest camp i mètriques semàntiques propostes en aquest treball. Mitjançant aquest estudi, es podran observar les diferències que existeixen entre les dues avaluacions comentades anteriorment i la millora que aporta la mesura semàntica proposta. Per a l’ús d’aproximacions de resum basades en xarxes neuronals i per a l’avaluació de tots els sistemes, es requereix un corpus de documents i resums. Per aquest motiu, mitjançant un procés de crawling, s’ha elaborat un corpus de notícies que provenen de distints llocs web de premsa digital i per als idiomes de castellà i català. Finalment, el treball conté un extens anàlisi dels resultats experimentals obtinguts en el corpus. En aquest es podran observar les diferències que existeixen entre les notícies que contenen ambdós corpus emprant distintes mesures. A més, es compararan els corpus generats amb altres corpus que formen part del estat de l’art., [EN] Automatic summary generation is a current and very attractive field in the area of natural language processing. For this reason, in this work, it has been decided to analyse and compare different techniques for automatic abstract generation both based on neural net-work and based on classical algorithms. For the evaluation of these techniques, syntactic metrics used in most studies of this field and semantic metrics proposed in this work have been used. Through this study, differences between the two evaluations discussed above can be seen. Moreover, improvements introduces by the semantic measure proposal are also presented. For the use of summary approaches based on neural networks and for the evaluation of all systems, a corpus of documents and summaries is required. For this reason, through a crawling process, a corpus of news that come from different digital press websites has been prepared in both Spanish and Catalan languages. Finally, the work contains an extensive analysis of the experimental results obtained with the corpus. In it, the existing differences between the news contained in both corpus can be observed. At the same time, a comparison of the different measures has been included. In addition, the generated corpus will be compared with other corpus that take part of the state of the art.

In recent years, the act of scraping websites for information has become increasingly relevant. However, along with this increase in interest, the internet has also grown substantially and advances and improvements to websites over the years have in fact made it more difficult to scrape. One key reason for this is that scrapers simply account for a significant portion of the traffic to many websites, and so developers often implement anti-scraping measures along with the Robots Exclusion Protocol (robots.txt) to try to stymie this traffic. The popular use of dynamically loaded content – content which loads after user interaction – poses another problem for scrapers. In this paper, we have researched what kinds of issues commonly occur when scraping and crawling websites – more specifically when scraping social media – and how to solve them. In order to understand these issues better and to test solutions, a literature review was performed and design and creation methods were used to develop a prototype scraper using the frameworks Scrapy and Selenium. We found that automating interaction with dynamic elements worked best to solve the problem of dynamically loaded content. We also theorize that having an artificial random delay when scraping and randomizing intervals between each visit to a website would counteract some of the anti-scraping measures. Another, smaller aspect of our research was the legality and ethicality of scraping. Further thoughts and comments on potential solutions to other issues have also been included.

In recent years, the act of scraping websites for information has become increasingly relevant. However, along with this increase in interest, the internet has also grown substantially and advances and improvements to websites over the years have in fact made it more difficult to scrape. One key reason for this is that scrapers simply account for a significant portion of the traffic to many websites, and so developers often implement anti-scraping measures along with the Robots Exclusion Protocol (robots.txt) to try to stymie this traffic. The popular use of dynamically loaded content – content which loads after user interaction – poses another problem for scrapers. In this paper, we have researched what kinds of issues commonly occur when scraping and crawling websites – more specifically when scraping social media – and how to solve them. In order to understand these issues better and to test solutions, a literature review was performed and design and creation methods were used to develop a prototype scraper using the frameworks Scrapy and Selenium. We found that automating interaction with dynamic elements worked best to solve the problem of dynamically loaded content. We also theorize that having an artificial random delay when scraping and randomizing intervals between each visit to a website would counteract some of the anti-scraping measures. Another, smaller aspect of our research was the legality and ethicality of scraping. Further thoughts and comments on potential solutions to other issues have also been included.

This study contributes novel digital mixed methods and findings on how fear of pedophilia affects parents and children’s bodily relations. We explore how norms for appropriate behavior between parents and children are constructed in the public debate on a specific case, where a mom has playful contact with her son’s genitals. The case triggered a public debate with both negative and positive reactions. A Laclau and Mouffe-inspired analytical framework and Internet-specific tools for data collection as well as processing contribute to the development of a new form of discourse analysis. This new discourse analysis is based on a combination of the digital tools word cloud and topic models, and a qualitative in-depth reading. By exploring discursive constructions and articulations of right and wrong, the study supports earlier findings that the online public unanimously agrees to the dictum of child innocence. However, we find openness within the discourses on how to define respectively healthy and damaging parental behavior towards children., This study contributes novel digital mixed methods and findings on how fear of pedophilia affects parents and children’s bodily relations. We explore how norms for appropriate behavior between parents and children are constructed in the public debate on a specific case, where a mom has playful contact with her son’s genitals. The case triggered a public debate with both negative and positive reactions. A Laclau and Mouffe-inspired analytical framework and Internet-specific tools for data collection as well as processing contribute to the development of a new form of discourse analysis. This new discourse analysis is based on a combination of the digital tools word cloud and topic models, and a qualitative in-depth reading. By exploring discursive constructions and articulations of right and wrong, the study supports earlier findings that the online public unanimously agrees to the dictum of child innocence. However, we find openness within the discourses on how to define respectively healthy and damaging parental behavior towards children.

[Resumen] El objetivo de este trabajo fin de grado es el desarrollo e implementación de un asistente virtual para el sitio web de la Facultad de Informática de la Universidad de A Coruña (UDC). Los usuarios serán capaces de interactuar con el asistente, tanto por entrada de texto como a través de voz, y le podrán pedir información sobre cualquier apartado de la web. Para promover el uso generalizado del asistente, se encontrará disponible tanto en gallego como en español. El proyecto seguirá una metodología que incluye todos los ciclos que corresponden a un proyecto software, desde un análisis previo para conocer los objetivos principales del proyecto, planificación, pasando por el diseño, implementación y pruebas. Con este trabajo se ha conseguido un sistema que implementa y demuestra las funcionalidades de un sistema conversacional., [Abstract] The objective of this end-of-degree project is to develop and implement a virtual assistant for the website of the College of Informatic Engineers of the University of A Coruña (UDC). Users will be able to interact with the assistant, both by keyboard inputs and voice, so they will be able to ask for any information of the different sections of the website. In accordance with the promotion of the generalized use of the assistant, its use will be available for both Galician and Spanish. The project will follow a methodology that includes all the software steps that correspond to a project with this characteristics, from a analysis that allows to know the main objectives of the projet, planning, through the design, implementation and finally testing. Most of the development was carried out following a framework that was planned in the initial phase of the project. With this work, it was achieved a system that implements and demonstrated the functionalities of a question answering process with an virtual assistant as its interface.

This study contributes novel digital mixed methods and findings on how fear of pedophilia affects parents and children’s bodily relations. We explore how norms for appropriate behavior between parents and children are constructed in the public debate on a specific case, where a mom has playful contact with her son’s genitals. The case triggered a public debate with both negative and positive reactions. A Laclau and Mouffe-inspired analytical framework and Internet-specific tools for data collection as well as processing contribute to the development of a new form of discourse analysis. This new discourse analysis is based on a combination of the digital tools word cloud and topic models, and a qualitative in-depth reading. By exploring discursive constructions and articulations of right and wrong, the study supports earlier findings that the online public unanimously agrees to the dictum of child innocence. However, we find openness within the discourses on how to define respectively healthy and damaging parental behavior towards children., This study contributes novel digital mixed methods and findings on how fear of pedophilia affects parents and children’s bodily relations. We explore how norms for appropriate behavior between parents and children are constructed in the public debate on a specific case, where a mom has playful contact with her son’s genitals. The case triggered a public debate with both negative and positive reactions. A Laclau and Mouffe-inspired analytical framework and Internet-specific tools for data collection as well as processing contribute to the development of a new form of discourse analysis. This new discourse analysis is based on a combination of the digital tools word cloud and topic models, and a qualitative in-depth reading. By exploring discursive constructions and articulations of right and wrong, the study supports earlier findings that the online public unanimously agrees to the dictum of child innocence. However, we find openness within the discourses on how to define respectively healthy and damaging parental behavior towards children.

Online social media, such as Facebook, Twitter, and LinkedIn, provides users with services that enable them to interact both globally and instantly. The nature of social media interactions follows a constantly growing pattern that requires selection mechanisms to find and analyze interesting data. These interactions on social media can then be modeled into interaction networks, which enable network-based and graph-based methods to model and understand users’ behaviors on social media. These methods could also benefit the field of complex networks in terms of finding initial seeds in the information cascade model. This thesis aims to investigate how to efficiently collect user-generated content and interactions from online social media sites. A novel method for data collection that is using an exploratory research, which includes prototyping, is presented, as part of the research results in this thesis. Analysis of social data requires data that covers all the interactions in a given domain, which has shown to be difficult to handle in previous work. An additional contribution from the research conducted is that a novel method of crawling that extracts all social interactions from Facebook is presented. Over the period of the last few years, we have collected 280 million posts from public pages on Facebook using this crawling method. The collected posts include 35 billion likes and 5 billion comments from 700 million users. The data collection is the largest research dataset of social interactions on Facebook, enabling further and more accurate research in the area of social network analysis. With the extracted data, it is possible to illustrate interactions between different users that do not necessarily have to be connected. Methods using the same data to identify and cluster different opinions in online communities have also been developed and evaluated. Furthermore, a proposed method is used and validated for finding appropriate seeds for information cascade

[ES] Se explica cómo se construye un catálogo unificador de la oferta hecha por tiendas en línea de cómics hentai digitales mediante las técnicas de rastreo (crawling) y scraping. Este catálogo está soportado por una plataforma que sigue la arquitectura orientada al servicio (SOA)., [EN] In this master's thesis it is explained how a unifying catalogue of the digital hentai comic books that are offered in online shops was created through crawling and scraping web techniques. This catalogue works in a service-oriented architecture (SOA) platform

[ES] Se explica cómo se construye un catálogo unificador de la oferta hecha por tiendas en línea de cómics hentai digitales mediante las técnicas de rastreo (crawling) y scraping. Este catálogo está soportado por una plataforma que sigue la arquitectura orientada al servicio (SOA)., [EN] In this master's thesis it is explained how a unifying catalogue of the digital hentai comic books that are offered in online shops was created through crawling and scraping web techniques. This catalogue works in a service-oriented architecture (SOA) platform

[ES] El incremento de la obesidad infantil ha suscitado, en los últimos años, el auge de técnicas de valoración de la actividad física, que es uno de los factores determinantes en la misma. Una de las técnicas punteras, en este campo, es la acelerometría. Este tipo de técnicas permiten la obtención de datos de intensidad, sin embargo, no permiten distinguir entre tipos de actividad. Los acelerómetros MEMS utilizados en estas técnicas, son capaces de recoger la información registrada y transmitirla a otro dispositivo, como una computadora. Con estos datos se pueden crear algoritmos de clasificación, que analizando los datos de entradas (aceleraciones), consiguen que a su salida se prediga el tipo de actividad que se está llevando a cabo. Bajo estas premisas, el objetivo del trabajo ha sido desarrollar un algoritmo capaz de clasificar las actividades infantiles de gateo, marcha y carrera, mediante las señales obtenidas a través de un acelerómetro triaxial. Para su creación se han empleado técnicas estadísticas cuya principal característica es la extracción de información de las señales obtenidas a través del acelerómetro. Una vez obtenida la información, estas técnicas también son capaces de averiguar la relación existente entre los datos. Para la realización del algoritmo, la técnica seleccionada ha sido la discriminante lineal. La misma predice como influyen cada una de las variables que diferencian a los distintos grupos que se han determinado para la clasificación. Durante la realización del trabajo se han extraído la media y la desviación típica en las tres direcciones del espacio (X, Y, Z), y después de examinar todos los datos, el modelo ha determinado los coeficientes de influencia de cada una de ellas en la marcha, gateo y carrera. Tras la creación del algoritmo, se ha utilizado una validación cruzada. En este tipo de validación los datos se dividen en dos grupos de datos, uno de ellos se utiliza para ajustar el modelo, y el otro para realizar un testeo. Cons, [CA] L'increment de l'obesitat infantil ha suscitat, en els últims anys, l'auge de tècniques de valoració de l'activitat física, que és un dels seus factors determinants. Una de les tècniques capdavanteres, en este camp, és l'acelerometría. Este tipus de tècniques permeten l'obtenció de dades d'intensitat, no obstant això, no permeten distingir entre els tipus d'activitat. Els acceleròmetres MEMS utilitzats en estes tècniques, són capaços de recollir la informació registrada i transmetre-la a un altre dispositiu, com una computadora. Amb estes dades es poden crear algorismes de classificació, que analitzant les dades d'entrades (acceleracions), aconsegueixen que a la seua eixida es prediga el tipus d'activitat que s'està duent a terme. Sota estes premisses, l'objectiu del treball ha sigut desenvolupar un algorisme capaç de classificar les activitats infantils del gateig, marxa i carrera, mitjançant els senyals obtinguts a través d'un acceleròmetre triaxial. Per a la seua creació s'han emprat tècniques estadístiques que tenen com a característica principal l'extracció d'informació dels senyals obtinguts a través de l'acceleròmetre. Una vegada obtinguda la informació, estes tècniques també són capaces d'esbrinar la relació existent entre les dades. Per a la realització de l'algorisme, la tècnica seleccionada ha sigut la discriminació lineal. Esta prediu com influeixen cadascuna de les variables que diferencien als diferents grups, que s'han determinat per a la classificació. Durant la realització del treball s'han extret la mitjana i la desviació típica en les tres adreces de l'espai (X, I, Z), i després d'examinar totes les dades, el model ha determinat els coeficients d'influència de cadascuna d'elles en la marxa, gateig i carrera. Després de la creació de l'algorisme, s'ha utilitzat una validació creuada. En este tipus de validació les dades es divideixen en dos grups de dades, un d'ells s'utilitza per ajustar el model, i l'altre per realitzar un testeig. Considera, [EN] In the recent years, the increasement in childhood obesity has cause the rise of techniques for the assessment of physical activity. This is due to the fact that physical activity is one of the determining factors in obesity. One of the leading techniques in this field is accelerometry. Although this technic allows obtaining intensity data, it does not allow to distinguish between types of activity. The MEMS accelerometers are used to solve this issue. These ones have the capacity of collecting the recorded information and transmitting it to another device, such as a computer. With this data, classification algorithms can be created. After its creation, they can analyze the input data (accelerations) and afterwards get a prediction, at its output, of the type of activity that is being carried out. Under these premises, the objective of these work has been to develop an algorithm capable of classifying children's activities, through the signals obtained through a triaxial accelerometer. The activities that were able to distinguish were: crawling, running and gait. For its creation, statistical techniques have been used. Their main characteristic is the capability of extracting information from the signals obtained through the accelerometer and then finding out the relationship between data. The selected technique for the work has been the linear discriminant analysis. This supervised method predicts how the variables selected for the classification can differentiate between groups and which is its influence. During the work, the mean and the standard deviation were extracted in the three directions of the space (X, Y, Z), and after examining all the data, the model was able to determine the coefficients of influence of each of three activities studied. After the creation of the algorithm, a cross-validation was used. In this type of validation, the data is divided into two groups, one of them is used to adjust the model, and the other is used for testing. Consid

[ES] Se explica cómo se construye un catálogo unificador de la oferta hecha por tiendas en línea de cómics hentai digitales mediante las técnicas de rastreo (crawling) y scraping. Este catálogo está soportado por una plataforma que sigue la arquitectura orientada al servicio (SOA)., [EN] In this master's thesis it is explained how a unifying catalogue of the digital hentai comic books that are offered in online shops was created through crawling and scraping web techniques. This catalogue works in a service-oriented architecture (SOA) platform

[ES] El incremento de la obesidad infantil ha suscitado, en los últimos años, el auge de técnicas de valoración de la actividad física, que es uno de los factores determinantes en la misma. Una de las técnicas punteras, en este campo, es la acelerometría. Este tipo de técnicas permiten la obtención de datos de intensidad, sin embargo, no permiten distinguir entre tipos de actividad. Los acelerómetros MEMS utilizados en estas técnicas, son capaces de recoger la información registrada y transmitirla a otro dispositivo, como una computadora. Con estos datos se pueden crear algoritmos de clasificación, que analizando los datos de entradas (aceleraciones), consiguen que a su salida se prediga el tipo de actividad que se está llevando a cabo. Bajo estas premisas, el objetivo del trabajo ha sido desarrollar un algoritmo capaz de clasificar las actividades infantiles de gateo, marcha y carrera, mediante las señales obtenidas a través de un acelerómetro triaxial. Para su creación se han empleado técnicas estadísticas cuya principal característica es la extracción de información de las señales obtenidas a través del acelerómetro. Una vez obtenida la información, estas técnicas también son capaces de averiguar la relación existente entre los datos. Para la realización del algoritmo, la técnica seleccionada ha sido la discriminante lineal. La misma predice como influyen cada una de las variables que diferencian a los distintos grupos que se han determinado para la clasificación. Durante la realización del trabajo se han extraído la media y la desviación típica en las tres direcciones del espacio (X, Y, Z), y después de examinar todos los datos, el modelo ha determinado los coeficientes de influencia de cada una de ellas en la marcha, gateo y carrera. Tras la creación del algoritmo, se ha utilizado una validación cruzada. En este tipo de validación los datos se dividen en dos grupos de datos, uno de ellos se utiliza para ajustar el modelo, y el otro para realizar un testeo. Cons, [CA] L'increment de l'obesitat infantil ha suscitat, en els últims anys, l'auge de tècniques de valoració de l'activitat física, que és un dels seus factors determinants. Una de les tècniques capdavanteres, en este camp, és l'acelerometría. Este tipus de tècniques permeten l'obtenció de dades d'intensitat, no obstant això, no permeten distingir entre els tipus d'activitat. Els acceleròmetres MEMS utilitzats en estes tècniques, són capaços de recollir la informació registrada i transmetre-la a un altre dispositiu, com una computadora. Amb estes dades es poden crear algorismes de classificació, que analitzant les dades d'entrades (acceleracions), aconsegueixen que a la seua eixida es prediga el tipus d'activitat que s'està duent a terme. Sota estes premisses, l'objectiu del treball ha sigut desenvolupar un algorisme capaç de classificar les activitats infantils del gateig, marxa i carrera, mitjançant els senyals obtinguts a través d'un acceleròmetre triaxial. Per a la seua creació s'han emprat tècniques estadístiques que tenen com a característica principal l'extracció d'informació dels senyals obtinguts a través de l'acceleròmetre. Una vegada obtinguda la informació, estes tècniques també són capaces d'esbrinar la relació existent entre les dades. Per a la realització de l'algorisme, la tècnica seleccionada ha sigut la discriminació lineal. Esta prediu com influeixen cadascuna de les variables que diferencien als diferents grups, que s'han determinat per a la classificació. Durant la realització del treball s'han extret la mitjana i la desviació típica en les tres adreces de l'espai (X, I, Z), i després d'examinar totes les dades, el model ha determinat els coeficients d'influència de cadascuna d'elles en la marxa, gateig i carrera. Després de la creació de l'algorisme, s'ha utilitzat una validació creuada. En este tipus de validació les dades es divideixen en dos grups de dades, un d'ells s'utilitza per ajustar el model, i l'altre per realitzar un testeig. Considera, [EN] In the recent years, the increasement in childhood obesity has cause the rise of techniques for the assessment of physical activity. This is due to the fact that physical activity is one of the determining factors in obesity. One of the leading techniques in this field is accelerometry. Although this technic allows obtaining intensity data, it does not allow to distinguish between types of activity. The MEMS accelerometers are used to solve this issue. These ones have the capacity of collecting the recorded information and transmitting it to another device, such as a computer. With this data, classification algorithms can be created. After its creation, they can analyze the input data (accelerations) and afterwards get a prediction, at its output, of the type of activity that is being carried out. Under these premises, the objective of these work has been to develop an algorithm capable of classifying children's activities, through the signals obtained through a triaxial accelerometer. The activities that were able to distinguish were: crawling, running and gait. For its creation, statistical techniques have been used. Their main characteristic is the capability of extracting information from the signals obtained through the accelerometer and then finding out the relationship between data. The selected technique for the work has been the linear discriminant analysis. This supervised method predicts how the variables selected for the classification can differentiate between groups and which is its influence. During the work, the mean and the standard deviation were extracted in the three directions of the space (X, Y, Z), and after examining all the data, the model was able to determine the coefficients of influence of each of three activities studied. After the creation of the algorithm, a cross-validation was used. In this type of validation, the data is divided into two groups, one of them is used to adjust the model, and the other is used for testing. Consid

Online social media, such as Facebook, Twitter, and LinkedIn, provides users with services that enable them to interact both globally and instantly. The nature of social media interactions follows a constantly growing pattern that requires selection mechanisms to find and analyze interesting data. These interactions on social media can then be modeled into interaction networks, which enable network-based and graph-based methods to model and understand users’ behaviors on social media. These methods could also benefit the field of complex networks in terms of finding initial seeds in the information cascade model. This thesis aims to investigate how to efficiently collect user-generated content and interactions from online social media sites. A novel method for data collection that is using an exploratory research, which includes prototyping, is presented, as part of the research results in this thesis. Analysis of social data requires data that covers all the interactions in a given domain, which has shown to be difficult to handle in previous work. An additional contribution from the research conducted is that a novel method of crawling that extracts all social interactions from Facebook is presented. Over the period of the last few years, we have collected 280 million posts from public pages on Facebook using this crawling method. The collected posts include 35 billion likes and 5 billion comments from 700 million users. The data collection is the largest research dataset of social interactions on Facebook, enabling further and more accurate research in the area of social network analysis. With the extracted data, it is possible to illustrate interactions between different users that do not necessarily have to be connected. Methods using the same data to identify and cluster different opinions in online communities have also been developed and evaluated. Furthermore, a proposed method is used and validated for finding appropriate seeds for information cascade

[ES] El incremento de la obesidad infantil ha suscitado, en los últimos años, el auge de técnicas de valoración de la actividad física, que es uno de los factores determinantes en la misma. Una de las técnicas punteras, en este campo, es la acelerometría. Este tipo de técnicas permiten la obtención de datos de intensidad, sin embargo, no permiten distinguir entre tipos de actividad. Los acelerómetros MEMS utilizados en estas técnicas, son capaces de recoger la información registrada y transmitirla a otro dispositivo, como una computadora. Con estos datos se pueden crear algoritmos de clasificación, que analizando los datos de entradas (aceleraciones), consiguen que a su salida se prediga el tipo de actividad que se está llevando a cabo. Bajo estas premisas, el objetivo del trabajo ha sido desarrollar un algoritmo capaz de clasificar las actividades infantiles de gateo, marcha y carrera, mediante las señales obtenidas a través de un acelerómetro triaxial. Para su creación se han empleado técnicas estadísticas cuya principal característica es la extracción de información de las señales obtenidas a través del acelerómetro. Una vez obtenida la información, estas técnicas también son capaces de averiguar la relación existente entre los datos. Para la realización del algoritmo, la técnica seleccionada ha sido la discriminante lineal. La misma predice como influyen cada una de las variables que diferencian a los distintos grupos que se han determinado para la clasificación. Durante la realización del trabajo se han extraído la media y la desviación típica en las tres direcciones del espacio (X, Y, Z), y después de examinar todos los datos, el modelo ha determinado los coeficientes de influencia de cada una de ellas en la marcha, gateo y carrera. Tras la creación del algoritmo, se ha utilizado una validación cruzada. En este tipo de validación los datos se dividen en dos grupos de datos, uno de ellos se utiliza para ajustar el modelo, y el otro para realizar un testeo. Cons, [CA] L'increment de l'obesitat infantil ha suscitat, en els últims anys, l'auge de tècniques de valoració de l'activitat física, que és un dels seus factors determinants. Una de les tècniques capdavanteres, en este camp, és l'acelerometría. Este tipus de tècniques permeten l'obtenció de dades d'intensitat, no obstant això, no permeten distingir entre els tipus d'activitat. Els acceleròmetres MEMS utilitzats en estes tècniques, són capaços de recollir la informació registrada i transmetre-la a un altre dispositiu, com una computadora. Amb estes dades es poden crear algorismes de classificació, que analitzant les dades d'entrades (acceleracions), aconsegueixen que a la seua eixida es prediga el tipus d'activitat que s'està duent a terme. Sota estes premisses, l'objectiu del treball ha sigut desenvolupar un algorisme capaç de classificar les activitats infantils del gateig, marxa i carrera, mitjançant els senyals obtinguts a través d'un acceleròmetre triaxial. Per a la seua creació s'han emprat tècniques estadístiques que tenen com a característica principal l'extracció d'informació dels senyals obtinguts a través de l'acceleròmetre. Una vegada obtinguda la informació, estes tècniques també són capaces d'esbrinar la relació existent entre les dades. Per a la realització de l'algorisme, la tècnica seleccionada ha sigut la discriminació lineal. Esta prediu com influeixen cadascuna de les variables que diferencien als diferents grups, que s'han determinat per a la classificació. Durant la realització del treball s'han extret la mitjana i la desviació típica en les tres adreces de l'espai (X, I, Z), i després d'examinar totes les dades, el model ha determinat els coeficients d'influència de cadascuna d'elles en la marxa, gateig i carrera. Després de la creació de l'algorisme, s'ha utilitzat una validació creuada. En este tipus de validació les dades es divideixen en dos grups de dades, un d'ells s'utilitza per ajustar el model, i l'altre per realitzar un testeig. Considera, [EN] In the recent years, the increasement in childhood obesity has cause the rise of techniques for the assessment of physical activity. This is due to the fact that physical activity is one of the determining factors in obesity. One of the leading techniques in this field is accelerometry. Although this technic allows obtaining intensity data, it does not allow to distinguish between types of activity. The MEMS accelerometers are used to solve this issue. These ones have the capacity of collecting the recorded information and transmitting it to another device, such as a computer. With this data, classification algorithms can be created. After its creation, they can analyze the input data (accelerations) and afterwards get a prediction, at its output, of the type of activity that is being carried out. Under these premises, the objective of these work has been to develop an algorithm capable of classifying children's activities, through the signals obtained through a triaxial accelerometer. The activities that were able to distinguish were: crawling, running and gait. For its creation, statistical techniques have been used. Their main characteristic is the capability of extracting information from the signals obtained through the accelerometer and then finding out the relationship between data. The selected technique for the work has been the linear discriminant analysis. This supervised method predicts how the variables selected for the classification can differentiate between groups and which is its influence. During the work, the mean and the standard deviation were extracted in the three directions of the space (X, Y, Z), and after examining all the data, the model was able to determine the coefficients of influence of each of three activities studied. After the creation of the algorithm, a cross-validation was used. In this type of validation, the data is divided into two groups, one of them is used to adjust the model, and the other is used for testing. Consid

Online social media, such as Facebook, Twitter, and LinkedIn, provides users with services that enable them to interact both globally and instantly. The nature of social media interactions follows a constantly growing pattern that requires selection mechanisms to find and analyze interesting data. These interactions on social media can then be modeled into interaction networks, which enable network-based and graph-based methods to model and understand users’ behaviors on social media. These methods could also benefit the field of complex networks in terms of finding initial seeds in the information cascade model. This thesis aims to investigate how to efficiently collect user-generated content and interactions from online social media sites. A novel method for data collection that is using an exploratory research, which includes prototyping, is presented, as part of the research results in this thesis. Analysis of social data requires data that covers all the interactions in a given domain, which has shown to be difficult to handle in previous work. An additional contribution from the research conducted is that a novel method of crawling that extracts all social interactions from Facebook is presented. Over the period of the last few years, we have collected 280 million posts from public pages on Facebook using this crawling method. The collected posts include 35 billion likes and 5 billion comments from 700 million users. The data collection is the largest research dataset of social interactions on Facebook, enabling further and more accurate research in the area of social network analysis. With the extracted data, it is possible to illustrate interactions between different users that do not necessarily have to be connected. Methods using the same data to identify and cluster different opinions in online communities have also been developed and evaluated. Furthermore, a proposed method is used and validated for finding appropriate seeds for information cascade

Online social media, such as Facebook, Twitter, and LinkedIn, provides users with services that enable them to interact both globally and instantly. The nature of social media interactions follows a constantly growing pattern that requires selection mechanisms to find and analyze interesting data. These interactions on social media can then be modeled into interaction networks, which enable network-based and graph-based methods to model and understand users’ behaviors on social media. These methods could also benefit the field of complex networks in terms of finding initial seeds in the information cascade model. This thesis aims to investigate how to efficiently collect user-generated content and interactions from online social media sites. A novel method for data collection that is using an exploratory research, which includes prototyping, is presented, as part of the research results in this thesis. Analysis of social data requires data that covers all the interactions in a given domain, which has shown to be difficult to handle in previous work. An additional contribution from the research conducted is that a novel method of crawling that extracts all social interactions from Facebook is presented. Over the period of the last few years, we have collected 280 million posts from public pages on Facebook using this crawling method. The collected posts include 35 billion likes and 5 billion comments from 700 million users. The data collection is the largest research dataset of social interactions on Facebook, enabling further and more accurate research in the area of social network analysis. With the extracted data, it is possible to illustrate interactions between different users that do not necessarily have to be connected. Methods using the same data to identify and cluster different opinions in online communities have also been developed and evaluated. Furthermore, a proposed method is used and validated for finding appropriate seeds for information cascade

[ES] El incremento de la obesidad infantil ha suscitado, en los últimos años, el auge de técnicas de valoración de la actividad física, que es uno de los factores determinantes en la misma. Una de las técnicas punteras, en este campo, es la acelerometría. Este tipo de técnicas permiten la obtención de datos de intensidad, sin embargo, no permiten distinguir entre tipos de actividad. Los acelerómetros MEMS utilizados en estas técnicas, son capaces de recoger la información registrada y transmitirla a otro dispositivo, como una computadora. Con estos datos se pueden crear algoritmos de clasificación, que analizando los datos de entradas (aceleraciones), consiguen que a su salida se prediga el tipo de actividad que se está llevando a cabo. Bajo estas premisas, el objetivo del trabajo ha sido desarrollar un algoritmo capaz de clasificar las actividades infantiles de gateo, marcha y carrera, mediante las señales obtenidas a través de un acelerómetro triaxial. Para su creación se han empleado técnicas estadísticas cuya principal característica es la extracción de información de las señales obtenidas a través del acelerómetro. Una vez obtenida la información, estas técnicas también son capaces de averiguar la relación existente entre los datos. Para la realización del algoritmo, la técnica seleccionada ha sido la discriminante lineal. La misma predice como influyen cada una de las variables que diferencian a los distintos grupos que se han determinado para la clasificación. Durante la realización del trabajo se han extraído la media y la desviación típica en las tres direcciones del espacio (X, Y, Z), y después de examinar todos los datos, el modelo ha determinado los coeficientes de influencia de cada una de ellas en la marcha, gateo y carrera. Tras la creación del algoritmo, se ha utilizado una validación cruzada. En este tipo de validación los datos se dividen en dos grupos de datos, uno de ellos se utiliza para ajustar el modelo, y el otro para realizar un testeo. Cons, [CA] L'increment de l'obesitat infantil ha suscitat, en els últims anys, l'auge de tècniques de valoració de l'activitat física, que és un dels seus factors determinants. Una de les tècniques capdavanteres, en este camp, és l'acelerometría. Este tipus de tècniques permeten l'obtenció de dades d'intensitat, no obstant això, no permeten distingir entre els tipus d'activitat. Els acceleròmetres MEMS utilitzats en estes tècniques, són capaços de recollir la informació registrada i transmetre-la a un altre dispositiu, com una computadora. Amb estes dades es poden crear algorismes de classificació, que analitzant les dades d'entrades (acceleracions), aconsegueixen que a la seua eixida es prediga el tipus d'activitat que s'està duent a terme. Sota estes premisses, l'objectiu del treball ha sigut desenvolupar un algorisme capaç de classificar les activitats infantils del gateig, marxa i carrera, mitjançant els senyals obtinguts a través d'un acceleròmetre triaxial. Per a la seua creació s'han emprat tècniques estadístiques que tenen com a característica principal l'extracció d'informació dels senyals obtinguts a través de l'acceleròmetre. Una vegada obtinguda la informació, estes tècniques també són capaces d'esbrinar la relació existent entre les dades. Per a la realització de l'algorisme, la tècnica seleccionada ha sigut la discriminació lineal. Esta prediu com influeixen cadascuna de les variables que diferencien als diferents grups, que s'han determinat per a la classificació. Durant la realització del treball s'han extret la mitjana i la desviació típica en les tres adreces de l'espai (X, I, Z), i després d'examinar totes les dades, el model ha determinat els coeficients d'influència de cadascuna d'elles en la marxa, gateig i carrera. Després de la creació de l'algorisme, s'ha utilitzat una validació creuada. En este tipus de validació les dades es divideixen en dos grups de dades, un d'ells s'utilitza per ajustar el model, i l'altre per realitzar un testeig. Considera, [EN] In the recent years, the increasement in childhood obesity has cause the rise of techniques for the assessment of physical activity. This is due to the fact that physical activity is one of the determining factors in obesity. One of the leading techniques in this field is accelerometry. Although this technic allows obtaining intensity data, it does not allow to distinguish between types of activity. The MEMS accelerometers are used to solve this issue. These ones have the capacity of collecting the recorded information and transmitting it to another device, such as a computer. With this data, classification algorithms can be created. After its creation, they can analyze the input data (accelerations) and afterwards get a prediction, at its output, of the type of activity that is being carried out. Under these premises, the objective of these work has been to develop an algorithm capable of classifying children's activities, through the signals obtained through a triaxial accelerometer. The activities that were able to distinguish were: crawling, running and gait. For its creation, statistical techniques have been used. Their main characteristic is the capability of extracting information from the signals obtained through the accelerometer and then finding out the relationship between data. The selected technique for the work has been the linear discriminant analysis. This supervised method predicts how the variables selected for the classification can differentiate between groups and which is its influence. During the work, the mean and the standard deviation were extracted in the three directions of the space (X, Y, Z), and after examining all the data, the model was able to determine the coefficients of influence of each of three activities studied. After the creation of the algorithm, a cross-validation was used. In this type of validation, the data is divided into two groups, one of them is used to adjust the model, and the other is used for testing. Consid

Online social media, such as Facebook, Twitter, and LinkedIn, provides users with services that enable them to interact both globally and instantly. The nature of social media interactions follows a constantly growing pattern that requires selection mechanisms to find and analyze interesting data. These interactions on social media can then be modeled into interaction networks, which enable network-based and graph-based methods to model and understand users’ behaviors on social media. These methods could also benefit the field of complex networks in terms of finding initial seeds in the information cascade model. This thesis aims to investigate how to efficiently collect user-generated content and interactions from online social media sites. A novel method for data collection that is using an exploratory research, which includes prototyping, is presented, as part of the research results in this thesis. Analysis of social data requires data that covers all the interactions in a given domain, which has shown to be difficult to handle in previous work. An additional contribution from the research conducted is that a novel method of crawling that extracts all social interactions from Facebook is presented. Over the period of the last few years, we have collected 280 million posts from public pages on Facebook using this crawling method. The collected posts include 35 billion likes and 5 billion comments from 700 million users. The data collection is the largest research dataset of social interactions on Facebook, enabling further and more accurate research in the area of social network analysis. With the extracted data, it is possible to illustrate interactions between different users that do not necessarily have to be connected. Methods using the same data to identify and cluster different opinions in online communities have also been developed and evaluated. Furthermore, a proposed method is used and validated for finding appropriate seeds for information cascade

With the growing use of popular social media services like Facebook and Twitter it is hard to collect all content from the networks without access to the core infrastructure or paying for it. Thus, if all content cannot be collected one must consider which data are of most importance.In this work we present a novel User-Guided Social Media Crawling method (USMC) that is able to collect data from social media, utilizing the wisdom of the crowd to decide the order in which user generated content should be collected, to cover as many user interactions as possible. USMC is validated by crawling 160 Facebook public pages, containing 368 million users and 1.3 billion interactions, and it is compared with two other crawling methods. The results show that it is possible to cover approximately 75% of the interactions on a Facebook page by sampling just 20% of its posts, and at the same time reduce the crawling time by 53%.What is more, the social network constructed from the 20% sample has more than 75% of the users and edges compared to the social network created from all posts, and has very similar degree distribution.

With the growing use of popular social media services like Facebook and Twitter it is hard to collect all content from the networks without access to the core infrastructure or paying for it. Thus, if all content cannot be collected one must consider which data are of most importance.In this work we present a novel User-Guided Social Media Crawling method (USMC) that is able to collect data from social media, utilizing the wisdom of the crowd to decide the order in which user generated content should be collected, to cover as many user interactions as possible. USMC is validated by crawling 160 Facebook public pages, containing 368 million users and 1.3 billion interactions, and it is compared with two other crawling methods. The results show that it is possible to cover approximately 75% of the interactions on a Facebook page by sampling just 20% of its posts, and at the same time reduce the crawling time by 53%.What is more, the social network constructed from the 20% sample has more than 75% of the users and edges compared to the social network created from all posts, and has very similar degree distribution.

[EN] HTML5 has several new components like XHR-Level2, DOM, Storage. With any major introduction of new features, HTML5 also brings with it potential security vulnerabilities. It allows crafting stealth attack vectors and adding risk to end client. Some originate from elements of the standard itself, some from implementations of the standard in each browser, and some from the care that developers do (or do not) take in building their HTML5 code. . . In this thesis we are going to talk about this new attack surface and possible threats. We are also going to cover how to automatically detect these possible vulnerabilities. This thesis describes a set of HTML5 vulnerabilities, we build a test web and use this for test the capabilities of several open source pen-testing tools. Based on the observed results, further work is also discussed., [ES] HTML5 tiene varios nuevos componentes como XHR-Level2, DOM, Storage. Con esta introducción de nuevas tecnologías, HTML5 también lleva consigo potenciales riesgos de seguridad. Algunos originados de los elementos del estándar en sí, otros de la implementación particular del estándar en cada navegador, y otros del cuidado que pongan los desarrolladores a la hora de escribir código. . . En esta tesis vamos hablas de estas nuevas estrategias de ataque y posibles amenazas. También cubriremos cómo detectar estas vulnerabilidad automatizando el proceso. Esta tesis describe una serie de vulnerabilidad web, sobre las que hemos construido unos test para probar las capacidad de algunas herramientas de pentesting. Basándonos en los resultados observados, discutiremos futuros resultados.

[EN] HTML5 has several new components like XHR-Level2, DOM, Storage. With any major introduction of new features, HTML5 also brings with it potential security vulnerabilities. It allows crafting stealth attack vectors and adding risk to end client. Some originate from elements of the standard itself, some from implementations of the standard in each browser, and some from the care that developers do (or do not) take in building their HTML5 code. . . In this thesis we are going to talk about this new attack surface and possible threats. We are also going to cover how to automatically detect these possible vulnerabilities. This thesis describes a set of HTML5 vulnerabilities, we build a test web and use this for test the capabilities of several open source pen-testing tools. Based on the observed results, further work is also discussed., [ES] HTML5 tiene varios nuevos componentes como XHR-Level2, DOM, Storage. Con esta introducción de nuevas tecnologías, HTML5 también lleva consigo potenciales riesgos de seguridad. Algunos originados de los elementos del estándar en sí, otros de la implementación particular del estándar en cada navegador, y otros del cuidado que pongan los desarrolladores a la hora de escribir código. . . En esta tesis vamos hablas de estas nuevas estrategias de ataque y posibles amenazas. También cubriremos cómo detectar estas vulnerabilidad automatizando el proceso. Esta tesis describe una serie de vulnerabilidad web, sobre las que hemos construido unos test para probar las capacidad de algunas herramientas de pentesting. Basándonos en los resultados observados, discutiremos futuros resultados.

[EN] HTML5 has several new components like XHR-Level2, DOM, Storage. With any major introduction of new features, HTML5 also brings with it potential security vulnerabilities. It allows crafting stealth attack vectors and adding risk to end client. Some originate from elements of the standard itself, some from implementations of the standard in each browser, and some from the care that developers do (or do not) take in building their HTML5 code. . . In this thesis we are going to talk about this new attack surface and possible threats. We are also going to cover how to automatically detect these possible vulnerabilities. This thesis describes a set of HTML5 vulnerabilities, we build a test web and use this for test the capabilities of several open source pen-testing tools. Based on the observed results, further work is also discussed., [ES] HTML5 tiene varios nuevos componentes como XHR-Level2, DOM, Storage. Con esta introducción de nuevas tecnologías, HTML5 también lleva consigo potenciales riesgos de seguridad. Algunos originados de los elementos del estándar en sí, otros de la implementación particular del estándar en cada navegador, y otros del cuidado que pongan los desarrolladores a la hora de escribir código. . . En esta tesis vamos hablas de estas nuevas estrategias de ataque y posibles amenazas. También cubriremos cómo detectar estas vulnerabilidad automatizando el proceso. Esta tesis describe una serie de vulnerabilidad web, sobre las que hemos construido unos test para probar las capacidad de algunas herramientas de pentesting. Basándonos en los resultados observados, discutiremos futuros resultados.

Heerema Marine Contractors (HMC) is a world leading marine contractor in the international offshore oil and gas industry. HMC specializes, among other activities, in the decommissioning of offshore located oil platforms. Current methods exist for safely removing topsides and substructures, yet research into new and cheaper ways is continuously done at HMC. Common practice now is to remove the topside with the use of a Semi-Submersible Crane Vessel (SSCV) and lift the topside onto a barge. With a new method, a reverse float-over operation, the SSCV would not be needed at all for this procedure. This operation has not been performed before. The set-up for this research is a barge which will connect to the topside due to friction forces on the interface between these two bodies. The use of stabbing cones or clamps would increase the horizontal station keeping of the barge beneath the topside during this operation. As this would require a lot of work being done on the topside before this operation can take place they are not used in this research. Therefore it was chosen to have the barge stick to the topside due to friction. In computer simulations a model of the H-851, a HMC owned barge, is located beneath an earth fixed topside. The deballasting of the barge is simulated by applying an external lift force on the centre of gravity of the barge. External wave forces are acting on the barge. When the barge is being pushed against the topside, friction forces allow the barge to crawl beneath the topside due to a stick slip mechanism. This thesis looks into the crawling behaviour by making a new friction model. A simplified model was created to improve the understanding of the friction forces between 2 bodies. A sensitivity study has been performed to analyse the crawling behaviour of the barge as a function of five parameters: Deballast velocity, vertical interface stiffness, horizontal interface stiffness, horizontal interface damping and coulomb fric

Heerema Marine Contractors (HMC) is a world leading marine contractor in the international offshore oil and gas industry. HMC specializes, among other activities, in the decommissioning of offshore located oil platforms. Current methods exist for safely removing topsides and substructures, yet research into new and cheaper ways is continuously done at HMC. Common practice now is to remove the topside with the use of a Semi-Submersible Crane Vessel (SSCV) and lift the topside onto a barge. With a new method, a reverse float-over operation, the SSCV would not be needed at all for this procedure. This operation has not been performed before. The set-up for this research is a barge which will connect to the topside due to friction forces on the interface between these two bodies. The use of stabbing cones or clamps would increase the horizontal station keeping of the barge beneath the topside during this operation. As this would require a lot of work being done on the topside before this operation can take place they are not used in this research. Therefore it was chosen to have the barge stick to the topside due to friction. In computer simulations a model of the H-851, a HMC owned barge, is located beneath an earth fixed topside. The deballasting of the barge is simulated by applying an external lift force on the centre of gravity of the barge. External wave forces are acting on the barge. When the barge is being pushed against the topside, friction forces allow the barge to crawl beneath the topside due to a stick slip mechanism. This thesis looks into the crawling behaviour by making a new friction model. A simplified model was created to improve the understanding of the friction forces between 2 bodies. A sensitivity study has been performed to analyse the crawling behaviour of the barge as a function of five parameters: Deballast velocity, vertical interface stiffness, horizontal interface stiffness, horizontal interface damping and coulomb fric

The Collection Management Webpages team is responsible for collecting, processing, and storing webpages from different sources. Our team worked on familiarizing ourselves with the necessary tools and data required to produce the specified output that was used by other teams in this class (Fall 2017 CS 5604). Input includes URLs generated by the Event Focused Crawler (EFC), URLs obtained from tweets by the Collection Management Tweets team, and webpage content from Web Archive (WARC) files from the Internet Archive or other sources. Our team fetches raw HTML from the obtained URLs and extracts HTML from WARC files. From this raw data, we obtain metadata information about the corresponding webpage. The raw data is also cleaned and processed for other teams' consumption. This processing is accomplished using various Python libraries. The cleaned information is made available in a variety of formats, including tokens, stemmed or lemmatized text, and text tagged with parts of speech. Both the raw and processed webpage data are stored in HBase and intermediately in HDFS (Hadoop Distributed File System). Our team successfully executed all individual portions of our proposed process. We successfully ran the EFC and obtained URLs from these runs. Using these URLs, we created WARC files. We obtained the raw HTML, extracted metadata information from it, and cleaned and processed the webpage information before uploading it to HBase. We iteratively expanded on the functionality of our cleaning and processing scripts in order to provide more relevant information to other groups. We processed and cleaned information from WARC files provided by the instructor in a similar manner. We have acquired webpage data from URLs obtained by the Collection Management Tweets (CMT) team. At this time however, there is no end-to-end process in place. Due to the volume of data our team has been dealing with, we explored various methods for parallelizing and speeding up our processes. Our team use