Search

Textual information is ubiquitous in our lives and is becoming an important component of our cognitive society. In the age of big data, we consistently need to traverse substantial amounts of data even to find a little information. To quickly acquire effective information, it is necessary to implement a textual similarity search based on an appropriate index structure to efficiently find results. In this article, we study top-k textual similarity search and develop a tree-based indexing approach that can construct indices to support various similarity functions. Our indexing approach clusters similar records in the same branch offline to improve the performance of online search. Based on the index tree, we present a top-k search algorithm with efficient pruning techniques. The experimental results demonstrate that our algorithm can achieve higher performance and better scalability than the baseline method.

Community search is the task of discovering dense subgraph that satisfy a set of given query parameters. Most community search algorithms consider link structure while ignoring link weight. A recent study proposed the idea of discovering weighted communities which focuses on both link structure and link weight using an online search approach and index-based approach. In this paper two online algorithms are proposed to scale-up the existing online approach efficiency. Performance evaluation of the proposed algorithms against the existing online approach over different datasets shows a great improvement in terms of search and query evaluation time.

These days most of the researchers use online search engines to find scientific articles they need. There are many specialized databases that are indexing scientific content and offer powerful search mechanism in these articles. On the other hand, authors and publishers are also interested of wide dissemination of their research articles and that is why they closely cooperate with these indexing organizations. In the paper are considered some popular indexing databases. We discuss their search functionality and practical aspects from authors and publishers’ point of view needed for content inclusion. Special attention is paid on the indexing mechanism of Google Scholar services.

Background: Most of the people are not medically qualified for studying or understanding the extremity of their diseases or symptoms. This is the place where natural language processing plays a vital role in healthcare. These chatbots collect patients' health data and depending on the data, these chatbot give more relevant data to patients regarding their body conditions and recommending further steps also. Purposes: In the medical field, AI powered healthcare chatbots are beneficial for assisting patients and guiding them in getting the most relevant assistance. Chatbots are more useful for online search that users or patients go through when patients want to know for their health symptoms. Methods: In this study, the health assistant system was developed using Dialogflow application programming interface (API) which is a Google's Natural language processing powered algorithm and the same is deployed on google assistant, telegram, slack, Facebook messenger, and website and mobile app. With this web application, a user can make health requests/queries via text message and might also get relevant health suggestions/recommendations through it. Results: This chatbot acts like an informative and conversational chatbot. This chatbot provides medical knowledge such as disease symptoms and treatments. Storing patients personal and medical information in a database for further analysis of the patients and patients get real time suggestions from doctors. Conclusion: In the healthcare sector AI-powered applications have seen a remarkable spike in recent days. This covid crisis changed the whole healthcare system upside down. So this NLP powered chatbot system reduced office waiting, saving money, time and energy. Patients might be getting medical knowledge and assisting ourselves within their own time and place.

Real-Time Bidding (RTB) is an important mechanism in online search advertising involved with the interaction among platform, advertisers, and customers, where the proper bid for each page views plays a crucial role for good marketing results. Enhanced Cost Per Click (ECPC) bidding is one of typical bidding strategies in the RTB mechanism where the advertisers will raise bid prices for traffic with high conversion rates under the premise of total advertisers' cost control in different time periods of the day. The bid ratio in ECPC as a correction coefficient is used to control the advertising cost of advertisers. However, the optimal bid ratio is hard to be derived due to the complexity and volatility of the auction environment. To address the challenges, we formulate the hourly-aggregated level of ECPC bidding as a Markov Decision Process with innovative state and reward function according to business. Besides, quite different from prior work that has to construct an auction model-based environment to solve the MDP problem, we propose a novel framework based on model-free reinforcement learning that does not rely on the agent's explorations in the environment. We furthermore design an adaptive action choosing strategy which adjusts the bidding behavior dynamically and improves the performance of categories/groups's consumption qualified rate. Experimental results on real dataset demonstrate the effectiveness of our framework from the perspective of consistency, stability and online effect.

The learning assistant method of the Physical-Cyber dual space is an effective method to improve learning interaction and learning effect. In this paper, we proposed a learning assistant method based on mixed reality technology and the fusion of physical space and cyber space, including two classical learning scenes. Firstly, we integrated physical real classroom scene with online three-dimensional resource scene, which effectively expanded the imagination of students. It allows students to learn abstract knowledge of macro or micro scenes more intuitively. Secondly, we explored the online search functions to extend resources related to offline teaching knowledge. Finally, we constructed the instructing scene and application example of the dual space fusion learning companion system using HoloLens. We have come to the necessity, rationality and technical feasibility of the method, and carry out learning activities on the basis of HoloLens, which provides new ideas for promoting the innovation of teaching and learning forms in the future.

In the era of modern technologies and with the development of marketing, nowadays the customer can save time buying things without having to visit the store. They just need to shop online, find what they are buying and pay it with just a few simple clicks. With this convenience in mind, the indulgence of people buying online has risen quite well. The customer’s decision to purchase a product through Internet commerce under the concept of electronic commerce - which has been developed into e-commerce or mobile commerce - usually takes a serious action with perceived confidence in the services and their suitability. Therefore, customer behaviour is an important factor to facilitate online electronic service. Hence, it is important to understand customer perceptions on these services and their intention of using it. Zalora is one of the fastest online search platforms providing kind of world-famous brands in Malaysia. The objective of this study is to examine factors that influence Malaysian consumers’ purchase intention through Zalora mobile application. Based on the theory of planned behaviour along with other factors, the study proposes a theoretical model for the empirical examination of the customer intention for purchasing via Zalora mobile application. Data were collected from sixteen respondents to test the relationship between variables. Data analysis was carried out using PLS-SEM. The empirical results indicated that perceived friendliness, perceived convenience, and perceived influence have a significant impact on purchase intention. The result may be preserved by providing an easy shopping experience to the user when using commerce sites and building trust to maintain the intent to buy. In addition, this analysis is helpful for Zalora marketers to suggest the most effective promotion methods in reference to further marketing of their products.

Graph similarity search is an important research problem in many applications, such as finding result graphs that have a similar structure to a given entity in biochemistry, data mining, and pattern recognition. Top-k graph similarity search is one of graph similarity search tasks, which aims to find the top-k graphs that are most similar to the query graph in a given graph database. In this paper, the top-k similarity search problem based on the graph edit distance is studied according to the corollary of the partitioned similarity theorem. Firstly, in order to speed up the online search process and avoid scanning each graph in the database one by one, an offline hierarchical inverted index is constructed to satisfy top-k search. Secondly, the offline hierarchical inverted index is used to filter the candidate graphs online and verify them, which reduces the time of searching the graphs. Finally, the good performance of the algorithm in running time and scalability is verified by running the similarity algorithm on real dataset and synthetic dataset.

The search behavior of Internet users is closely related to the needs of users, which is very important for the decision-making of enterprises. Based on the online search behavior, this paper makes an empirical study on the prediction of product sales, and establishes a model to reflect customer willingness. The findings show that there is a positive correlation between keyword search frequency and product sales. The contribution of this paper is to provide a quantitative model between search behavior and product sales, which can help enterprises to predict product sales and make better decisions.

Information Retrieval (IR) is the process of retrieving relevant documents from a large collection of documents. The applications of these technologies are no longer restricted to only online search engines, instead, these are widely being used by organizations to facilitate different organizational information searches. Efficient Retrieval of documents from a large collection of organizational data is still a challenging task. In this paper, a domain categorization-based Information Retrieval system is presented for offline document repository. Several tests were conducted to evaluate the performance of the presented model. The performance of the model is also compared with existing retrieval techniques and the results obtained show the efficiency of the model in providing relevant documents quickly. The model achieves approximately 93% precision and 92% recall accuracy in the considered domains and data set.

The proliferation of e-commerce websites in recent times have spurred the number of online reviews on products generated online. This rapid development of ecommerce is usually accompanied by the desire of consumers to search for more information on a product before making any purchase. These two activities undertaken by consumers who tend purchase products online, generates large volumes of data that provides some useful insight for product manufacturers. The data generated from online reviews and consumers online search activity on a product could help product manufacturers to determine the needs and requirements of consumers. Using data mining methods, this study proposes a methodology to estimate the value of the importance of product attributes. These product attributes are the customers’ needs and requirements mined from online reviews and consumers’ online search data. This study proposes to use Shapley value and Choquet integral to determine the importance of product attributes from online big data.

With the emergence of Big Data, the exploitation of information has become such a direct and rapid action that many questions arise about the reliability and value of data. However, the data raised by online search engines is based on a varied and massive environment, which includes varied data and from different sources of information. Indeed, in the field of distance education (e-Learning) and in particular during online research, the learner is faced with a massive mixture of data which does not exist in an ordered format and which is not easy to operate. This variety of data represents the second V in Big Data. In order to support learners in their search for information, it is proposed to design a specialized tool to combine these massive and varied data so that they can provide the appropriate result to the learner. This article deals with the descriptive study of the current state of the use of online research in the Big Data environment by the students of the UAE (Abdelmalek Essaadi University), and to highlight the advantages and impacts of this phenomenon on learning and scientific research.

The focus of this research is to examine the usage patterns exhibited by users of online search engines in the midst of COVID-19. We aim to understand how the queries are structured and their timing on the various platforms that citizens are using to check the availability of Personal Protective Equipment (PPE) since the outbreak of the COVID-19 public health crisis. Understanding and analyzing peak volume for information platforms is critical, especially for public health policy, with a mind toward crisis informatics. In this study, we collect all the data of users querying data from Face Mask Map (FMM), a real-time application which displays the inventory status for all stores selling PPE. This data is from the point at which the public health crisis became widely known to the time at which PPE availability saturated the market. As COVID-19 continues to proliferate and affect people around the globe, official organizations such as Department of Health and World Health Organization (WHO) utilize Web or Social Media (Facebook or Twitter) to announce up-to-date news, e.g. daily confirmed cases or in order to update policy regarding resource management. We then correlate the significant announcements from public health officials, specifically published by Ministry of Health and Welfare (MoHW) in Taiwan, that are concerning usage and distribution of PPE. We find that the temporal dynamics of aggregated users behavior are consistent with the events. For the practitioner of disaster management, it is critical to be able to identify when the public will consistently react to public health announcements for the purpose of ensuring proper supply distribution and avoid misallocation. It is our hope that the study can help to build an effective online disaster preparedness information system, in the consideration of computing and public psychology, to better respond to disaster with a greater corpus of data.

Searching homologous vulnerabilities in firmware from different platforms is key for protecting the Internet of Things (IoT) devices. Due to the hard work to acquire the source code, the existing works mainly focused on binary code, and many methods have been used to find vulnerabilities in binary code, like 1) numerical features are used to achieve efficient prefiltering, however, a large number of experiments showed that this method is which is lack of adaptability in some platforms; 2) structural features are used to achieve the accurate matching, but it will bring low efficiency. Based on the shortcomings of the existing methods, we proposed a lightweight method to search vulnerabilities of firmware in a cross-platform model. The main idea of this method is to exploit readable string literals inside functions combined with local call graph and use these string literals as a feature of each function in form of text, then MinHash LSH and MinHash LSH Forest algorithms are used to achieve similarity calculation of this feature according to different online search scenarios. Finally, we employed an optimization stage based on numeric features to improve the rank of similar function, which allows us to search for a similar function more accurately. The experimental results on the open library OpenSSL showed that our method has more robust performance in terms of filtering accuracy, and it is about 1 order of magnitude faster than the two existing methods in terms of efficiency.

Searchable Encryption (SE) provides data online search function to users while protecting the data security and the user's query privacy. It has challenges of high computational and storage overhead that should be mitigated for a practical SE system. In this paper, we propose a blockchain-based user management scheme for SE system that can decentralize bandwidth resource consumption. Meanwhile it prevents single point failure and enhances the security protection in order to prevent identity leakage, loss, tampering brought by TTP. It leverages blockchain to offer non-central dynamic user management capability. Our theoretical analysis validates the feasibility and security. Further, experiments are carried out to demonstrate the SE effectiveness.

Based on the Baidu search index of key words, this paper builds online search indexes of keywords related to Beijing tourism concerning the six basic elements of one trip– catering, accommodation, traffic, sightseeing, shopping and entertainment. A vector autoregression model is then built with the time serial date of the online search indexes and the actual number of Beijing domestic trips to forecast the latter. It is found that there is a long-term co-integration relationship between the online search indexes data and the number of domestic trips to Beijing. The model constructed has a good prediction ability for Beijing domestic trips.

The structural diversity of an edge, which is measured by the number of connected components of the edge’s ego-network, has recently been recognized as a key metric for analyzing social influence and information diffusion in social networks. Given this, an important problem in social network analysis is to identify top-k edges that have the highest structural diversities. In this work, we for the first time perform a systematical study for the top-k edge structural diversity search problem on large graphs. Specifically, we first develop a new online search framework with two basic upper-bounding rules to efficiently solve this problem. Then, we propose a new index structure using near-linear space to process the top-k edge structural diversity search in near-optimal time. To create such an index structure, we devise an efficient algorithm based on an interesting connection between our problem and the 4-clique enumeration problem. In addition, we also propose efficient index maintenance techniques to handle dynamic graphs. The results of extensive experiments on five large real-life datasets demonstrate the efficiency, scalability, and effectiveness of our algorithms.

Industry 4.0 is on the horizon. Therefore, it is crucial to analyze the patterns and trends of intellectual property (IP) information to determine the readiness of stakeholders to adapt to the changing industrial evolution. Patent bibliography documents consist of structured and unstructured data, so text mining or machine learning must be employed for the data analysis. This paper established a patent trend by analyzing the patent data of Intellectual Property Corporations of Malaysia (MyIPO) to identify the institution’s readiness to face the fourth industrial revolution. To achieve this aim, a patent classification method was used to classify MyIPO patent data based on the pillars of Industry 4.0. Furthermore, the patents data were drawn from MyIPO Online Search and Filing System was used as the datasets in this study. However, the dataset consists of the title of the patent and the publication year only. Since short text data in the title has fewer semantic information and high sparseness, this issue was a challenge for this study. In this paper, five common classifiers were used for text classification. Support Vector Machine (SVM) was proven to be the machine learning classifier with the highest accuracy in classifying the training and testing datasets. The findings of this paper present the patent trend for each pillar of Industry 4.0 including the patents related to Industry 4.0 where Autonomous Robot is the pillar with the highest innovation.

There are many incurable diseases around us like various types of cancers which are detected at the last stage. But they are curable if early detection is possible. Web search log with symptoms can be effective for such early detection. Online search queries can act as a useful evidence of user's health condition as they search with symptoms in search engine. In this paper, we have used machine learning techniques to identify those attributes from online search log that can be used for early detection of liver cancer. Liver cancer is diagnosed very late most of the time. Primary stage symptoms are obscure and vague in most of the cases. It becomes very difficult to get the treatment properly because of the metastasis of the cancer when diagnosed. We have proposed a method to organize online search log data to detect liver cancer early. If we have online search log data containing symptoms searched on search engine and the time of the searches of each symptom, they can be easily organized by our method of dataset creation and detect liver cancer. We have estimated the performance of Support Vector Machine (SVM) based on different kernel scale values. The best classification accuracy for $\gamma=0.03$ is 94.30%. We have also compared the performance of different machine learning techniques. Among them, Random Forest has the highest classification accuracy of 97.50%.

Online search and trading algorithms are designed with the prime objective of maximizing the overall profit of investors in a financial market. Investors in any financial market tend to manage their risk levels for better overall returns by the end of investment horizon. Thus the real world applicability of such algorithms lacks the aspect of risk management policy. We consider the problem of online k-min search, where an online player (investor) wants to to acquire (buy) $k$ units of an asset by using the minimum possible wealth. In this work, we investigate the k-min search problem and introduce the risk aware k-min search algorithm under the risk reward framework. Further, we show that the presented algorithm meets the bounds presented by AI-Binali (Algorithmica, 1999).

We propose a planning method to enable fast autonomous flight in cluttered environments. Typically, autonomous navigation through a complex environment requires a continuous search on a graph generated by a k-connected grid or a probabilistic scheme. As the vehicle travels, updating the graph with data from onboard sensors is expensive as is the search on the graph especially if the paths must be kinodynamically feasible. We propose to avoid the online search to reduce the computational complexity. Our method models the environment differently in two separate regions. Obstacles are considered to be deterministically known within the sensor range and probabilistically known beyond the sensor range. Instead of searching for the path with the lowest cost (typically the shortest path), the method maximizes the likelihood to reach the goal in determining the immediate next step for navigation. With such a problem formulation, the online method realized by a trajectory library can determine a path within 0.2-0.3ms using a single CPU thread on a modem embedded computer. In experiments, it enables a lightweight UAV to fly at 10m/s in a cluttered forest environment (see Fig. 1 as an example).

Previous research has proven that online search data can be used to estimate tourist visits. Massive query data is a challenge in determining the right keywords to used to build indexes. This study proposes a framework for building a composite search index. The proposed forecasting framework emphasizes keyword selection using keywords in previous research and is based on the tourism organization's website. Previous studies predict the number of tourist arrivals at a tourist place, city or city-state, whereas in this study predict foreign tourist visits at the country level, Indonesia. Using the econometric model, this study uses data on Foreign Tourist Visits (FTV), Google Trends Index (GTI), Customer Price Index (CPI) and Exchange Rate (ER). The forecasting method uses the Vector Error Correction Model (VECM) and then analyzes the model, provides forecasting and structural analysis of the model. The results of the model analysis are long-term and short-term analyzes and the results of forecasting evaluation show that the MAPE score is good.

Sales forecasting is an essential element for implementing sustainable business strategies in the automotive industry. Accurate sales forecasts enhance the competitive edge of car manufacturers in the effort to optimize their production planning processes. We propose a forecasting technique that combines keyword-specific customer online search data with economic variables to predict monthly car sales. To isolate online search data related to pre-purchase information search, we follow a backward induction approach and identify those keywords that are frequently applied by search engine users. In a set of experiments using real-world sales data and Google Trends, we find that our keyword-specific forecasting technique reduces the out-of-sample error by 5% as compared to existing techniques without systematic keyword selection. We also find that our regression models outperform the benchmark model by an out-of-sample prediction accuracy of up to 27%.

Programmers face situations where they have to rely on messy documentation, other developers and online search for basic programming commands and queries when they encounter any new programming environment. This leads to the waste of time of developers and decreases productivity. In this paper, we present, “PACT”, a chat bot which assists the programmers with basic programming queries that they face when they are new to a programming environment. We use Neural Machine Translation architecture to generate coherent, non-rule based responses to a programmer’s query. The data that is fed to the neural machine translation model is collected from websites like StackOverflow, technical sub-reddits and technical StackExchanges.

Marketing by using diverse distribution and communication channels or commonly called Omnichannel is an approach that integrates many marketing channels to create an effective and consistent consumer experience. The variety of marketing channels used by consumers, they expect to receive good service, the same shopping experience and are integrated in all types of distribution. This study aims to find out the profile of consumers of shopping online sites in Jakarta and find out any variables from Theory of Planned Behavior (TPB) which describes the online search behavior of omnichannel consumers. In testing the hypothesis in this study used SEM or Structural Equation Model. Respondents in this study were online retail consumers in Jakarta with 125 people. From the results of the study, it is obtained data that the profile is young women under the age of 30 years with the work of workers and employers are consumers of online retail shopping sites in Jakarta. By using the variables in the TPB model, it is known that consumers’ online buying behavior is influenced by the intention to purchase the product where this intention is formed by the presence of variable attitudes towards behavior, subjective norms and perceptions.

Recommender systems aim to provide personalized suggestions to customers which products to buy or services to consume. They can help to increase sales by helping customers discover new and relevant products. Traditionally, recommender systems use the purchase history of a customer, e.g., the purchased quantity or properties of the items. While this allows to build personalized recommendations, it is a very limited view of the problem. Nowadays, extensive information about customers and their personal preferences is available which goes far beyond their purchase behaviour. For example, customers reveal their preferences in social media, by their browsing habits and online search behaviour or their interest in specific newsletters. In this paper, we investigate how information from different sources and channels can be collected and incorporated into the recommendation process. We demonstrate this, based on a real-life case study of a retailer with several million transactions. We discuss how to employ a recommender system in this scenario, evaluate various recommendation strategies and describe how to incorporate information from different sources and channels, both internal and external. Our results show that the recommendations can be better tailored to the personal preferences of customers.

A large number of people die around the world in consequence of a fall. The costs related to fatal and nonfatal falls have an enormous impact on society and have been growing over the years. There are several risk factors that increase the probability of falling, such as poor balance and lower extremity weakness. Patients with balance impairment, in order to overcome these problems, use walkers. The aim of this work is to do an analysis of the fall-related strategies already implemented in a smart walker. Therefore, an online search was performed based on the literature through Scopus and Web of Science databases. A study was also conducted on a commercial level on Google, as well as a patent review on Espacenet and United States Trademark Office. It was possible to conclude that exist a concern related to the development of an approach to prevent the fall event. However, the only implemented strategy that was found throughout this research consists in stopping the walker when a near fall is detected.

Augmenting a robot with the capacity to understand the activities of the people it collaborates with in order to then label and segment those activities allows the robot to generate an efficient and safe plan for performing its own actions. In this work, we introduce an online activity segmentation algorithm that can detect activity segments by processing a partial trajectory. We model the transitions through activities as a hidden Markov model, which runs online by implementing an efficient particle-filtering approach to infer the maximum a posteriori estimate of the activity sequence. This process is complemented by an online search process to refine activity segments using task model information about the partial order of activities. We evaluated our algorithm by comparing its performance to two state-of-the-art activity segmentation algorithms on three human activity datasets. The proposed algorithm improved activity segmentation accuracy across all three datasets compared with the other two approaches, with a range from 11.3% to 65.5%, and could accurately recognize an activity through observation alone for 31.6% of the initial trajectory of that activity, on average. We also implemented the algorithm onto an industrial mobile robot during an automotive assembly task in which the robot tracked a human worker’s progress and provided the worker with the correct materials at the appropriate time.

In business and entrepreneurship, patents are used to gain a competitive advantage. Accordingly, engineers and educators should understand basic rules and standard operating procedures relating to patents. In research and development, patents may be used as sources of scientific literature. Patents could help identify technology landscapes and the state-of-the-art in a specific field. Importantly, patents can help inform product and manufacturing design by identifying what has already been done, and what is possible for future improvements and innovations. Accordingly, engineers and educators should understand how to use patents in their R&D work. From a career perspective, knowledge of patents and how to find them can distinguish individuals from their peers and aid them in going beyond the classroom or the lab. In this workshop, attendees will learn some basics about patents, and gain some hands on experience in how to search and identify patents, in order to accelerate their research or design projects. Attendees can expect an interactive workshop where they will learn patent basics and apply knowledge by searching patents using powerful free online search tools.

An online search approach to perform cooperative actions among simulated soccer players have shown practical performance. However, it is still not easy to design an appropriate evaluation function for the search process. In order to solve this problem, this paper proposes a method that applies learning to rank and a method to compose training data with smaller operational cost using clickthrough data. Experimental result shows our method successfully improved the performance of simulated soccer players in the small subtask.

Using seasonal topics as the study subject, in this study, we focus on the timing gap between social media writing and online search behavior. To conduct our analysis, we used the mathematical model of search behavior, comprising the sociophysics approach. The seasonal topics selected were cherry-blossom viewing for spring, bikinis for summer, autumn leaves for fall, and skiing for winter. We also picked up the event like Christmas and Halloween. We analyzed the influence of blogs and Twitter on search behavior and found a deviation of interest in terms of timing. We also analyzed seasonal topics in newspapers after 2010 and observed deviations in the number of searches and spikes in the number of references in newspaper magazines.

In software development, there is a great demand for online information and resources. The traditional way for the developers to access online resources is through formulating keywords and searching in the web browser. The search results are limited by the keywords and the web browser also ignores the developers' working and search context. Tools that integrate information retrieval into the IDE are available, but they fail to perceive the developers' dynamic working context and use in the process of online search. In this paper, we present a context-aware program assistant called amAssist. amAssist monitors the developers' development events and models their working context dynamically, then integrates them with the entire online search process~(e.g. keywords formulation, customized searching, search results annotation, etc.). We integrate amAssist into the Eclipse IDE. Our preliminary user study showed that by using our program assistant, developers can formulate keywords more accurately and acquire online information and resources more rapidly. Demo video: https://youtu.be/X4Tkjhc6wfU

With growing demands on location-based services, the indoor localization has attracted great attention in both academia and industry. This paper focuses on improving indoor localization performance based on Wi-Fi signal processing. Specifically, we propose an annulus local search based localization (ALSL) algorithm, which combines the distance fitting and fingerprint-based techniques. First, ALSL derives the relationship between the Received Signal Strength Indicator (RSSI) and the Euclidean distance for a Wi-Fi Access Point (AP) based on multinomial fitting. Then, an annulus construction scheme is proposed to reduce the online search space for fingerprint-based localization. On this basis, a probability model is derived to estimate the probability of an online point (OP) which is located at certain reference point (RP) based on Bayes theorem. Finally, K-Nearest-Neighbor (KNN) algorithm is adopted to compute the final estimated location of an OP. For performance evaluation, we build a prototype of the indoor localization system and implement the ALSL algorithm. We conduct a series of experiments in real-world scenarios, which show that ALSL is able to effectively reduce the computational overhead during the online localization phase with the annulus based local search policy. In addition, our field testing also demonstrates that ALSL can outperform existing competitive solutions in terms of improving the localization accuracy.

The purpose of this study is to build a web-based customized Corporate Identity System, with a search engine to facilitate logo searching, suggest logos and assist in the design processes. This web-based system should be icon-oriented, based on a database of corporate logos, logotypes and applicable product modules. With the assistance of such a system, designers and clients can generate a customized design process. They can find the logo, logotype and the supplementary design elements from the databases and apply them to business items, banners, packaging, uniforms, transportation and so on, using the online search engine. From this integrated platform, the basic elements of Corporate Identity Design can be modularized, edited and presented immediately on the internet, providing designers and their clients with the most time efficient means of communicating and testing ideas, Collective Intelligence becomes prevalent, enterprises are now facing even more rapid changes on the internet. Aside from getting hold of the core technology, effectively utilizing the synergy brought by Collective Intelligence, selecting the most appropriate corporate image solutions, and creating an information sharing center are crucial actions; it is also necessary to accumulate and share knowledge, and elevate the value of corporate image and culture so that the competitive advantages can be sustained.

Mobile devices have been an indispensable part of human life, which enable people to search and browse what they want on the move. Mobile video search, as one of the most important services for users, still faces great challenges under mobile internet scenario, such as the limitation of computation ability, memory, and bandwidth. Therefore, this paper proposes a multi-modal hash based framework for instant mobile video search. In particular, we adopt a efficient deep convolutional neural network, MobileNet, with the hash layer to learn discriminative and compact visual features from videos. Moreover, we also consider hand-crafted local visual descriptor and audio fingerprint to build a multi-modal hash representation of videos. With the multi-modal hash code, two types of hash indexes are built on the server to achieve efficient video search. At last, the multi-modal hash codes are extracted on the mobile devices and transferred in a three- step progressive procedure during the online search stage. The experiments on the real-world dataset show that the proposed framework not only achieves the state-of-the-art accuracy but also obtains excellent efficiency.

Search control in an unknown space is a basic function of mobile autonomous robots. For search control, robustness against environmental change is important because environmental information is measured and updated at every moment. Robustness can be guaranteed by feedback controllers generated by global control Lyapunov functions (CLFs). To endow search control with the robustness properties of CLFs, Akiba et al. proposed search control based on a Lyapunov technique in an unknown space using a Q-learning algorithm. They showed that the robot reaches its destination with conventional control even if the destination is unknown. However, conventional control requires an offline calculation because the Q-learning algorithm has high calculation costs; hence, an online search is not achieved. Further, conventional control cannot be applied to complex environments such as those where humans live. In this research, we propose a Lyapunov based on search control in an unknown plane space. The proposed control employs Dijkstra's algorithm, and has lower calculation costs than conventional search control. Further, we successfully apply our method to mapping data in real environments via a simulation experiment.

Finding approximated shortest paths for extremely large-scale complex networks is a challenging problem, where existing works require large overhead to achieve high accuracy and diversity for estimated paths, especially for large graphs with millions of vertices. In this paper, we propose an online search approach based on preprocessed indexes, to approximate point-to-point shortest paths. The approach is able to find more accurate and diverse paths with limited index overhead and requires low search overhead. Furthermore, a new path degree based index construction algorithm is introduced that can greatly increase the approximation accuracy and involve no additional index overhead. To handle extreme size graphs, we build a query processing system with our algorithm on distributed graph processing platforms. The system also supports parallel processing of online searches to achieve high throughput for a large number of queries. We evaluate our algorithm on various real-world graphs from different disciplines with up to billions of edges, and we demonstrate that our system can process hundreds of thousand queries per second on these graphs with reducedoverhead.

A new high efficient hierarchical pathfinding algorithm named Floyd-A∗(FA∗) is presented based on grid map of indoor environment for a mobile service robot. In FA∗, an algorithm for key points extraction and region partition is designed according to distinction of indoor environment so that the representation of abstract map about indoor environment is more effective for high speed pathfinding. In offline search phase, Floyd-Warshall algorithm is used to acquire some useful knowledge as priori ones so that a more efficient query operation can be conducted to replace the online search process on abstract maps. Moreover, a kind of matrix memory mechanism is employed to greatly reduce memory overheads and improve the capacity of saving abstract paths. A series of experiments and evaluation analysis on benchmarks show that FA∗ speeds up A∗ by one to two orders of magnitude and makes significant improvement on efficiency over some other excellent variants of A∗ on grid map so as to drastically enhance the performance of our proposed pathfinding approach for further practical applications.

Tourists usually do an online search of tourist destinations before determining their destinations. The Internet provides media to share information and experiences on tourism through electronic communication channels, such as social media, travel blogs and tourist review sites. The tourism sector has a strong influence from social media in the decision-making process and has big potential for promotions. However, it is difficult to filter the vast amount of information on the Internet. It takes a lot of time to be able to read all the reviews that are on the Internet, but reading only few reviews lead to biased evaluation. Therefore, we proposed an application that can perform sentiment analysis on travel destination reviews, especially in Indonesia.

The usage of online search engine is growing rapidly not only in daily life, but also in education. We are interested in understanding what strategies students apply during search, especially the tactics they use to decompose a programming task. In this paper, we report a lab study to investigate students' programming information seeking behavior via Google search engine. Students were given a programming task with limited time, and they were also required to report their online search process including search query and web pages browsed. We analyze the web pages they browsed, model student's behavior, and cluster them into groups with different search tactics. The results show that the web pages they browsed during the task consisted of either conceptual knowledge or coding technical content. The students who performed better would browsed more about conceptual knowledge. Students who set more and smaller unit of subgoals outperformed the students with fewer and larger subgoals.

Seasonal influenza is a contagious respiratory illness that can cause various complications, worsen chronic illnesses, and sometimes lead to deaths. During 2009 H1N1 flu pandemic, up to 203,000 deaths occurred worldwide. Early detection and prediction of disease outbreak is critical because it can provide more time to prepare a response and significantly reduce the impact caused by a pandemic. The traditional influenza surveillance system by Centers for Disease Control and Prevention (CDC) collects U.S. Influenza-Like Illness related physicians visits data from sentinel practices and provides a retrospective analysis delayed by two weeks. Google Flu Trends proposed a method that uses online search queries data to estimate current (real-time) influenza activity. Here we present a system that (1) predicts future influenza activities, (2) provides more accurate real-time assessment than before, and (3) combines real-time big social media data streams and CDC historical datasets for predictive models to accomplish accurate predictions. Although retrospective analysis and observations are important, prediction of future flu levels can represent a big leap because such predictions provide actionable insights for public health that can be used for planning, resource allocation, treatments and prevention. Thus, compared to previous work, our work represents an advancement in accuracy of assessments, prediction of future flu activity accurately and an ability to combine big social data and observed CDC data to build predictive models.

Big Data, true to its name, it deals with large volumes of data characterized by volume, variety and velocity. Big data has made the development of highly capable online search engines nowadays. Search Engine systems are differing by the way of how Indexing and Page Rankings are performed. Without Indexing, Search Engine would require considerable time and computing power. For example, while an index of 10,000 documents can be queried within milliseconds, a sequential scan of every word in 10,000 large documents could take hours. The common indexing methods are not suitable for Search Engine Big Data as it greatly increase the size of the data as well as reduce the scalability. This paper proposes a new method of Indexing Search Engine Big Data called Key Hash Indexing scheme followed by the implementation of Page Rank. A Comprehensive presentation of important technology and factors to achieve efficient Big Data storage, Indexing and Ranking in Web Search Engine are also considered. This system also shows the efficiency of the method with an extensive set of experiments on real data. Experimental results on real-time data sets show that the proposed solution is effective as well as efficient in index generation and ranking.

As an improvement of classical PageRank, the personalized PageRank soon became one of the most major ranking algorithm in graph computation. However, it suffers from a severe efficiency issue and there are many studies focus on enhancing its precision and lowering down its complexity, among which the Mento Carlo random approximation estimation performs well in time and the power iteration method handles with accuracy better. In this paper, several optimization methods are combined to make it practical for personalized social network searching. It leverages the localization property of social network with local searching and quick dot product computing for the sparse matrix. It also optimizes the searching performance by pre-computing the intermediate data and parallelizing for the real-time computation. The experimental result shows that our improved personalized PageRank works well in the individual online search of the social network which can meet the online search time requirement with a practical storage usage.

This work describes ongoing research on web information retrieval using queries which require prolonged browser interactions in exploring, revising, resubmitting, and studying results produced by the search engine. Over the last two decades or so, web search has improved significantly in a lot of areas, but has not made much progress at the difficult end of the online search spectrum — one that is precisely characterized by queries with complex information needs. Such queries, we call learning queries, require extended browser interaction by the user for satisfactory results. The main goal of the research is to address the difficulty of learning queries by lessening browser interactions through the implementation of a probabilistic graphical model. The expected outcome is an improvement in relevance, inspired by automatic discovery of collaborative search structures in conjunction with leveraging current search engine capabilities. To attain the project goal, we design and implement a browser extension that collects post-query browsing details based on browser interactions by users. The post-query browsing details are used to create a database of query-indexed trees for validating the model proposed. The work presented here describes the design and implementation of browser extensions and the project methodology.

Web of world facilitates search results on massive collection of data using Internet. While obtaining search results corresponding to a particular search query, authentication of search results becomes an important question, as different sources contribute data in every moment which are stored in (un)trusted third party servers. In this paper, we discuss on the authenticated search results of some recent works and then present an improved scheme that ensures the authenticity of the search results corresponding to a search query over Internet. The improved scheme is based on the scheme [1] that uses the concept of conjunctive keyword search. We first observe some weaknesses in [1] and then construct an improved scheme using strong authentication primitives. We show that the improved scheme can achieve better efficiency and mitigate the security weaknesses in [1]. The prototype implementation of the improved scheme shows its practicality in future Internet applications including online search engines.

As information technology has evolved, digital media has become increasingly fragmented and has started to proliferate multiple information channels. In order to optimize on the various digital channels that are available, organizations are increasingly recognizing the importance of gaining solid insights into consumer behavior and preferences that can be translated into marketing strategies. Specifically, they are keen to identify which information channels they can use to effectively reach and communicate with their target market. This paper describes how fuzzy AHP and TOPSIS can be used to develop a new method of decision making that will enable an effective and systematic decision process. This paper provides a demonstration of the underpinning working methodology of the proposed model by examining an illustrative example that is based on the decision process Internet users employ during their online search for information.

With the increasing of data, traditional centralized searching engine can not provide efficient searching service in the face of vast amounts of data. Aiming at this problem, cloud computing platform is used to quickly handle huge amounts of data. Hadoop cloud computing platform is used for distributed index creation. Central node is responsible to receive user's searching request and distribute them to searching nodes. After search nodes complete searching tasks, the searching results are merged by central node, which are finally sent to the user. Besides, cache technology is adopted that effectively reduces the load of search engine. The experiment results show this proposed platform has good searching performance.

The global objective of this study is to propose solutions for improving the quality of online search services. We consider the special case where the processing of requests submitted to such services consists of: the querying of several types of sub-services followed by the composition of the output produced the sub-services. This could be the case for instance of a booking service that proposes travel packs including: hotel reservation, car rental and flight booking. We propose to improve the quality of such services with the multi-selection problem. The goal in this problem is to select the subset of sub-services of each type to use in the the processing of search request. The selection must ensure that we maximize the quality of the results we could expect from the search. The multi-selection problem is close to the service selection problem. However, while in the latter, we are interested in a unique sub-service per type, in the former, we want to choose a subset of sub-services. Our paper introduces a theoretical formulation of the problem and demonstrates its NP-hardness. We also propose two approaches for the resolution. The first approach is based on Integer Linear Programming. The second approach combines parallel algorithm portfolio and sampling techniques. Finally, we make a comparative evaluation of the approaches based on practical scenarios.

Technological innovation has given rise to many internet portals including those that facilitate matrimony. Adoption of this technology has accelerated the growth of online matrimony portals in India. There are several sources of search for proposals and online matrimony is one such medium. The importance given to such portals has impacted the way matrimonial information is available in these portals. This study focuses on understanding factors that impact the importance given to online matrimony portals by users over other traditional ways. A conceptual model was developed and tested empirically with pan India survey. The sample size included 133. The data was analyzed using ordered logistic regression. This study identified that age, facilitating comparison, time taken, parents' level of education, community beliefs all significantly impact the perceived of importance given to online matrimony portals.