Search

Natural language processing is used for solving a wide variety of problems. Some scholars and interest groups working with language resources are not well versed in programming, so there is a need for a good graphical framework that allows users to quickly design and test natural language processing pipelines without the need for programming. The existing frameworks do not satisfy all the requirements for such a tool. We, therefore, propose a new framework that provides a simple way for its users to build language processing pipelines. It also allows a simple programming language agnostic way for adding new modules, which will help the adoption by natural language processing developers and researchers. The main parts of the proposed framework consist of (a) a pluggable Docker-based architecture, (b) a general data model, and (c) APIs description along with the graphical user interface. The proposed design is being used for implementation of a new natural language processing framework, called ANGLEr.

V nalogi smo za obdobje 1976%2005 preučevali odzive telesnih kazalnikov jelenjadi glede na številčnost jelenjadi, ki je določenega leta poseljevala območje LPN Jelen Sneţnik. Telesni kazalci, ki smo jih analizirali, so telesna masa ločeno po spolu in starostnih kategorijah za vse osebke, pri odraslih jelenih smo analizirali še maso rogovja in število CIC točk rogovja. Populacijsko gostoto smo rekonstruirali s pomočjo povratnega računanja iz podatkov, dobljenih iz evidenc odvzete jelenjadi. Rezultati so pokazali, da se je številčnost populacije gibala od 5,8 pa do 2,6 osebka na 100 hektarjev. Analize kaţejo, da populacijska gostota ni statistično značilno vplivala na telesno maso telet in enoletne jelenjadi. Pokazal pa se je vpliv populacijske gostote na telesno maso odrasle jelenjadi in to pri obeh spolih. Pri analizi vplivov na maso rogovja in število CIC točk rogovja smo ugotovili, da nanju pozitivno vplivata višja starost in višja telesna masa, negativno pa nanju vpliva populacijska gostota. In this graduation thesis we were analyzing body indicators of red deer in the area of LPN Jelen Sneţnik, depending to population density in the period from 1976 to 2005. Analyzed body indicators are: body mass separated by sex and age category for all individuals, for adult male we analyzed also antler mass and CIC points of antler. Population density was reconstructed using reverse calculation, from data obtained from year harvesting. The results showed that the population density varied from 2,6 up to 5,8 individuals per 100 hectares. Analysis show that population density do not statistically significant influence the body weight of deer calves and yearlings. On the other hand we found statistically significant influence of population density on body weight for both sex of older animals. In analyzing the effects on antler weight and number of antler CIC points, we found a positive influence on them by older age and higher body weight, but they were influenced negatively by population density.

[EN] Mutation Testing is a well-established technique for assessing the quality of test cases by checking how well they detect faults injected into a software artefact (mutant). Using this technique, the most critical activity is the adequate design of mutation operators so that they reflect typical defects of the artefact under test. This paper presents the design of a set of mutation operators for Conceptual Schemas (CS) based on UML Class Diagrams (CD). In this paper, the operators are defined in accordance with an existing defects classification for UML CS and relevant elements identified from the UML-CD meta-model. The operators are subsequently used to generate first order mutants for a CS under test. Finally, in order to analyse the usefulness of the mutation operators, we measure some basic characteristics of mutation operators with three different CSs under test., This work has been developed with the financial support by SENESCYT of the Republic of Ecuador, European Commission (CaaS project) and Generalitat Valenciana (PROMETEOII/2014/039)

International audience; Observing information systems projects shows that information systems engineering (ISE) methods are underused. The iAMF project aims at (a) tracing stakeholders' activities to identify whether this statement is true and (b) proposing more efficient ISE methods. To trace stakeholders' activities, we need a tool able to record any computerized actions - as opening applications, modifying documents, compiling programs, etc. This paper presents a review of trace-based tools that was undertaken to address the issue of recording information systems engineering methods enactment. We followed the MADISE decision making approach to select the most appropriate trace-based tool for the iAMF project.

Organization’s concerns as controlling costs, improving quality, increasing effectiveness, and managing risk increasingly impose strong requirements on Business-IT alignment (B-ITa). Several Maturity Models came into being for considering improvement actions in B-ITa. The IT-Enabled Collaborative Network organizations (ICoNOs) maturity model specifically addresses the processes needed for achieving B-ITa in networked organizations. We have evaluated ICoNOs maturity model on a case study in an Asian organization. On the basis of this study, we propose to introduce a new domain named as “Cost Management‿.

Complex networks of real-world systems are believed to be controlled by common phenomena, producing structures far from regular or random. Clustering, community structure and assortative mixing by degree are perhaps among most prominent examples of the latter. Although generally accepted for social networks, these properties only partially explain the structure of other networks. We first show that degree-corrected clustering is in contrast to standard definition highly assortative. Yet interesting on its own, we further note that non-social networks contain connected regions with very low clustering. Hence, the structure of real-world networks is beyond communities. We here investigate the concept of functional modules---groups of regularly equivalent nodes---and show that such structures could explain for the properties observed in non-social networks. Real-world networks might be composed of functional modules that are overlaid by communities. We support the latter by proposing a simple network model that generates scale-free small-world networks with tunable clustering and degree mixing. Model has a natural interpretation in many real-world networks, while it also gives insights into an adequate community extraction framework. We also present an algorithm for detection of arbitrary structural modules without any prior knowledge. Algorithm is shown to be superior to state-of-the-art, while application to real-world networks reveals well supported composites of different structural modules that are consistent with the underlying systems. Clear functional modules are identified in all types of networks including social. Our findings thus expose functional modules as another key ingredient of complex real-world networks., Comment: 21 pages, 15 figures, 7 tables

Community structure is largely regarded as an intrinsic property of complex real-world networks. However, recent studies reveal that networks comprise even more sophisticated modules than classical cohesive communities. More precisely, real-world networks can also be naturally partitioned according to common patterns of connections between the nodes. Recently, a propagation based algorithm has been proposed for the detection of arbitrary network modules. We here advance the latter with a more adequate community modeling based on network clustering. The resulting algorithm is evaluated on various synthetic benchmark networks and random graphs. It is shown to be comparable to current state-of-the-art algorithms, however, in contrast to other approaches, it does not require some prior knowledge of the true community structure. To demonstrate its generality, we further employ the proposed algorithm for community detection in different unipartite and bipartite real-world networks, for generalized community detection and also predictive data clustering.

Fraud in motor insurance is assessed to incur annual losses in the range of 100 billion dollars. While much research exists in the fraud management field, majority only deals with partial problems and presupposes the independence of specific fraud management activities. Researches on components of fraud management system are rarely explicitly related to business performance improvements. These results in a common problem, which can be observed on the practitioners’ side: only small amount of companies can objectively assess which of the many fraud management system components proposed by researchers and vendors will help to solve their problems in fraud management. The method proposed in this paper can be used as a strategic tool for improvement of fraud management process in motor insurance companies. The method is designed to be used for a selection of fraud management system components, and is based on business performance. The input for the method is a set of key performance indicators that an insurance companies wish to improve. The result is a set of activities, which should be improved, and a set of fraud management system components that should be used to improve these activities. The paper presents and explains the method and its components. The method components have been developed based on the data received from Slovenian motor insurance companies and method is evaluated in three case studies. Article in English. Transporto priemonių draudimo apgavysčių valdymo sistemos komponentų pasirinkimo metodas, grindžiamas verslo veiklos efektyvumu Santrauka.Apgavystės transporto priemonių draudimo srityje lemia draudimo kompanijų nuostolius, kurie vertinami 100 milijardų dolerių per metus. Dėl to apgavysčių valdymo srityje atliekama nemažai tyrimų, tačiau daugumoje iš jų nagrinėjamos tik dalinės problemos ir daromos prielaidos, kad apgavysčių valdymo veiklos rūšys nepriklauso viena nuo kitos. Apgavysčių valdymo sistemų komponentų tyrimai retai tiesiogiai siejami su jų įtaka verslo veiklos efektyvumui. Tai lemia dažną problemą, kuri būdinga praktiniam apgavysčių valdymo sistemų naudojimui: tik nedaug draudimo kompanijų gali objektyviai įvertinti, kuris iš daugelio apgavysčiu valdymo sistemų komponentų, siūlomų tyrėjų ir pardavėjų, padėtu jiems išspręsti problemas, susijusias su apgavysčių valdymu. Darbe siūlomas metodas, kuris gali būti panaudotas kaip strateginė priemonė, skirta apgavysčių valdymo procesui patobulinti transporto priemonių draudimo kompanijose. Metodas sukurtas taip, kad galėtų būti naudojamas apgavysčių valdymo sistemu komponentams pasirinkti remiantis verslo veiklos efektyvumu. Pradiniai metodo naudojami duomenys yra svarbiausių veiklos efektyvumo rodiklių, kuriuos kompanijos nori patobulinti, rinkinys. Rezultatas yra veiklos rūšių, kurios turėtų būti patobulintos, rinkinys ir apgavysčių valdymo sistemos komponentų rinkinys, kuris turėtų būti panaudotas sprendžiant suformuluotą apgavysčių valdymo uždavinį. Straipsnyje aprašytas pasiūlytas metodas ir siūlomi naudoti sistemos komponentai. Metodą realizuojantys komponentai buvo sukurti remiantis duomenimis, kurie gauti iš Slovėnijos transporto priemonių draudimo kompanijų. Pasiūlytas metodas ir jį įgyvendinanti sistema buvo išbandyti atliekant tris praktinius eksperimentus, kurių rezultatai pateikti straipsnyje. Reikšminiai žodžiai: apgavysčių valdymas,apgavysčių valdymo sistemos,apgavysčių valdymo sistemų kūrimas,verslo veiklos efektyvumas,svarbiausi veiklos efektyvumo rodykliai,transporto priemonių draudimas.

Label propagation has proven to be a fast method for detecting communities in complex networks. Recent work has also improved the accuracy and stability of the basic algorithm, however, a general approach is still an open issue. We propose different label propagation algorithms that convey two unique strategies of community formation, namely, defensive preservation and offensive expansion of communities. Furthermore, the strategies are combined in an advanced label propagation algorithm that retains the advantages of both approaches; and are enhanced with hierarchical community extraction, prominent for the use on larger networks. The proposed algorithms were empirically evaluated on different benchmarks networks with planted partition and on over 30 real-world networks of various types and sizes. The results confirm the adequacy of the propositions and give promising grounds for future analysis of (large) complex networks. Nevertheless, the main contribution of this work is in showing that different types of networks (with different topological properties) favor different strategies of community formation.

Situational method engineering (SME) is concerned with the design, construction, and evaluation of methods that are aimed at the development of particular software products in well-defined development situations. The SME approach features characteristics such as flexibility, adaptability, modularity, reusability, and reference to situational aspects. These aspects may not only be useful for the engineering of software but also for the engineering of work systems, organizational structures, or enterprises. This paper therefore presents exemplary applications of the SME approach in domains that are different from software engineering. It furthermore discusses potentials and limitations of the transfer and application of SME principles, concepts, and techniques to those subject areas.

V diplomski nalogi se ukvarjam z izdelavo strežnika za obdelavo podatkov, ki jih pošiljajo Arduino mikrokrmilniki. Arduino krmilniki merijo vlago v rastlinah in nato te meritve pošiljajo na strežnik. Strežnik meritve zbere, jih obdela ter robotu posreduje pot in lokacije rastlin, ki jih mora zaliti. V prvem delu so opisani različni pristopi za reševanje problema trgovskega potnika in tudi opis uporabljene rešitve. Ta rešitev je kasneje uporabljena za računanje poti, ki jo mora opraviti robot. V nadaljevanju so predstavljene tehnologije in protokoli, ki so bili uporabljeni za izdelavo strežnika, podatkovne baze, čelnega dela sistema in prenos podatkov. Sledi opis izdelave in delovanja strežnika ter implementacija podatkovne baze. In my thesis, I deal with the creation of a data processing server, sent by Arduino microcontrollers. Arduino controllers measure humidity in plants, and then send these measurements to the server. The server collects these metrics processes them and provides the robot with the path and locations of the plants it needs to water. The őrst part describes different approaches to solving the traveling salesperson problem and also describes the solution used. This solution is later used for calculating the path that the robot must take. The next section presents the technologies and protocols used to create the server, database, front end of the system, and data transfer. At the end, there is a description of the development and operation of the server and database implementation.

Mesto in urbanizacija ponujata veliko prednosti za ljudi, kot je centraliziranost storitev in dejavnosti, vendar lahko hkrati povzročata tudi negativne posledice, kot so zgoščen promet, hrup in onesnaževanje zraka. Osebna vozila so glavni krivec za onesnaževanje zraka v mestih, zato nekatere države Evropske unije uvajajo prepoved uporabe avtomobilov na bencinski in dizelski pogon. Alternativa temu so električna vozila, vendar si jih mnogi še vedno ne morejo privoščiti. Koncept skupne uporabe električnih vozil oz. car sharing predstavlja udobno in okolju prijazno obliko transporta, ki omogoča uporabo avtomobila brez njegovega lastništva in znižuje mesečne stroške. Ta koncept je v Sloveniji podprt s strani Ministrstva za infrastrukturo in Slovenske platforme za trajnostno mobilnost. Storitve car sharing so uporabniku prijazne in lahko dostopne, saj se izposoja izvaja preko aplikacije. Podatki, ki jih aplikacija in izposojevalne postaje zbirajo, so lahko uporabni za načrtovanje javne prometne infrastrukture. Internet stvari omogoča povezovanje teh naprav z ostalimi pametnimi napravami v mestu. Cilj naše naloge je s pomočjo tehnologij FIWARE umestiti storitev car sharing v okvir pametnih mest preko spletne aplikacije za pregled zasedenosti vozil te storitve. Končni produkt diplomskega dela bo torej spletni pregledovalnik zasedenosti vozil car sharing, podatki pa bodo shranjeni na gradniku FIWARE Context Broker, ki predstavlja osrednjo komponento naše aplikacije. S tem bomo preverili ustreznost komponent FIWARE za hranjenje, obdelavo in prikaz podatkov postaj car sharing. The city and urbanization offer many advantages for people, such as centralization of services and activities, but they can also have negative consequences, such as concentrated traffic, noise and air pollution. Private cars are a major cause of urban air pollution, which is why some EU countries are introducing bans on petrol and diesel cars. Electric vehicles are an alternative, but many people still can't afford them. The concept of electric car sharing is a convenient and environmentally friendly form of transport that allows people to use a car without owning it and reduces monthly costs. This concept is supported in Slovenia by the Ministry of Infrastructure and the Slovenian Platform for Sustainable Mobility. Car sharing services are user-friendly and easily accessible, as car rental is done via an app. The data collected by the app and the car rental stations can be useful for public transport infrastructure planning. The Internet of Things (IoT) makes it possible to connect these devices with other smart devices in the city. The aim of our thesis is to use FIWARE technologies to place car sharing services in the smart cities framework through a web application to check the occupancy of the vehicles of this service. The final product of the thesis will therefore be a web-based occupancy viewer for car sharing vehicles, and the data will be stored on the FIWARE widget Context Broker, which is the central component of our application. This will verify the suitability of the FIWARE components for storing, processing, and displaying car sharing station data.

V diplomski nalogi smo se posvetili problemu hranjenja domačih živali na daljavo. Predstavili smo obstoječi postopek hranjenja s časovnikom na fiksen časovni interval. Opisali smo način doziranja s pomočjo dozirnega vijaka in načrtovali ohišje, potrebno za umestitev naših komponent. Predstavili smo vso strojno opremo, uporabljeno pri izdelavi hranilnika, ter načrtovali sistem, v katerem s pomočjo mobilne aplikacije kontroliramo ročno ali avtomatsko hranjenje. Avtomatsko sprožanje je nadzorovano s poljubnim številom urnikov ponavljanja, ki jih upravljamo z mobilno aplikacijo. Naš hranilnik podpira tudi različne vrste hranjenj, za katera lahko določimo poljubne količine in imena. Te podatke upravljamo s pomočjo zalednega sistema, ki nato podatke shranjuje v podatkovni bazi PostgreSQL. In this thesis, we have focused on the problem of feeding pets remotely. We presented an existing feeding procedure via the timer at a fixed time interval. We described the method of dosing using a dosing screw and designed the necessary housing to store our components. We presented all the hardware used for the implementation of the feeder and designed a system in which manual or automatic feeding is controlled with mobile application. Automatic triggering is controlled by any number of management schedules, which we manage via the mobile app. Our feeder also supports different types of feeding, for which we can define quantities and names. This data is managed by a back-end system, which then stores the data on a PostgreSQL database.

Diplomska naloga se osredotoča na zasnovo avtomatskega testiranja programske opreme, ki se je v zadnjih letih uveljavilo kot najbolj učinkovita metoda za odkrivanje napak in zagotavljanje optimalnega delovanja. Zato je postalo avtomatsko testiranje nepogrešljivo orodje v svetu razvoja programske opreme, saj je zasenčilo staromodno ročno testiranje. V nadaljevanju naslavlja izzive, s katerimi se srečujemo pri testiranju aplikacij z govornim vmesnikom ter predstavi ustrezna orodja in ogrodja za avtomatsko testiranje tovrstnih aplikacij. Med naštetimi sta nato izbrani dve orodji, in sicer Playwright ter Selenium, ki ju konkretno predstavimo in primerjamo. Glavni cilj diplomske naloge je tako preveriti orodja za avtomatsko testiranje aplikacij z govornim vmesnikom. Poudarek je na enostavnosti, učinkovitosti ter hitrosti pisanja in izvajanja testov. Končni rezultat dela nam pove, katero orodje je najbolj primerno za testiranje aplikacij, ki zadostujejo našim potrebam. This thesis focuses on the concept of automated software testing, which has emerged in recent years as the most effective method for detecting faults and ensuring optimal performance. As a result, automated testing has become an indispensable tool in the world of software development, eclipsing old-fashioned manual testing. In the following, we address the challenges encountered when testing applications with a speech interface and present relevant tools and frameworks for automated testing of such applications. Two of these tools, Playwright and Selenium, are then selected for concrete presentation and comparison. The main objective of the thesis is thus to examine tools for automatic testing of speech-enabled applications. The focus is on simplicity, efficiency and speed of writing and running tests. The final result of the work tells us which tool is the most suitable for testing applications that meet our needs.

Law and court documents may contain sensitive information that should not be accessible to everyone. Most of the publicly available documents are only partially anonymized or use a general anonymization system that makes the once anonymized data unsuitable for further research. We focus on developing several types of anonymization systems, with implementing NER, POS tagging, coreference resolution as well as specific anonymization rules that target Slovene entities. The anonymization models are integrated into a publicly available web tool, which allows convenient and fast anonymization in four different modes, depending on the sensitivity of the documents. The obtained results prove that the developed models can successfully anonymize Slovene documents. The anonymized documents can be distributed without violating someone's privacy and can be used in various research in the future. Pravni in sodni dokumenti lahko vsebujejo občutljive podatke, ki ne bi smeli biti dostopni vsem. Večina javno dostopnih dokumentov je le delno anonimiziranih ali pa uporablja splošni sistem anonimizacije, zaradi česar so enkrat anonimizirani podatki neprimerni za nadaljnje raziskovanje. Osredotočeni smo na razvoj več tipov anonimizacijskih sistemov, z implementacijo NER, POS označevanja, koreferenčne ločljivosti ter specifičnih anonimizacijskih pravil, ki ciljajo na slovenske subjekte. Modeli anonimizacije so integrirani v javno dostopno spletno orodje, ki omogoča priročno in hitro anonimizacijo v štirih različnih načinih, odvisno od občutljivosti dokumentov. Dobljeni rezultati dokazujejo, da lahko razviti modeli uspešno anonimizirajo slovenske dokumente. Anonimizirane dokumente je mogoče distribuirati brez poseganja v zasebnost nekoga in jih je mogoče uporabiti v različnih raziskavah v prihodnosti.

Glasovni asistenti postajajo vse bolj priljubljeni in uporabni, vendar nobeden ne podpira slovenščine. Prav tako nismo odkrili pomočnikov, ki bi podpirali celoten cevovod glasovnega asistenta za slovenščino. V diplomskem delu smo razvili glasovnega asistenta za slovenski jezik za androidno napravo. Prepoznava ključne besede poteka lokalno na napravi, brez odvisnosti od interneta, s prilagojenim akustičnim modelom. Prepoznava govora poteka na zunanji storitvi z večjim akustičnim modelom. Za logiko glasovnega asistenta smo uporabili storitev Rasa, ki iz transkripcije uporabnikovega govora prepozna namen in kontekst. Odgovor te storitve se uporabi za sintezo odgovora, ki se nato predvaja uporabniku. Logiko na storitvi Rasa smo razširili, tako da ta storitev podpira tudi dva nova scenarija. Aplikacijo smo testirali v tihem okolju in v okolju s šumom za oba spola. Testirali pa smo tudi, delovanje aplikacije z uporabo druge storitve za transkripcijo govora. Aplikacija je najbolje delovala v tihem okolju za moškega govorca, saj v tem primeru ni prišlo do izpada iz toka delovanja aplikacije in smo čez celoten cevovod prišli v 100% testiranj. Primerljivo dobro pa je delovala tudi z uporabo druge storitve za transkripcijo govora, pri kateri smo čez celoten cevovod prišli v 94% testiranj. Voice assistants are becoming more and more popular and useful. However, none of the best-known voice assistants support the Slovene language. We also have not found any assistants that would support the entire pipeline of the voice assistant for Slovene language. We have developed a voice assistant for Slovene language and for an Android device. Keyword detection is done locally on the device, without needing the Internet, and with a modified acoustic model. For speech recognition, we used an outside service with a bigger language model. For the logic of the voice assistant, we used Rasa. Rasa can detect the intent and context of the transcribed speech. The response from this service is used for speech synthesis and is then played back to the user. We tested the application in two different environments, silent and noisy, and for the male and female voice. We also tested if the application functions properly if we use another service for speech recognition. The application worked best in the silent environment for a male speaker. In that case we finished the pipeline in 100% of tests. We achieved comparable results when we used another service for speech recognition. We finished the pipeline in 94% of tests.

Včasih se znajdemo v situaciji, ko je pisanje na roko ali tipkanje na tipkovnico nepraktično, zamudno ali je zaradi posameznikovih motoričnih ovir oteženo. V takšnih primerih je dobrodošla rešitev možnost uporabe glasovnega narekovanja. V diplomski nalogi smo se ukvarjali z razvojem narekovalnika za slovenski jezik in njegovo integracijo v Microsoft Word. Natančneje, razvili smo dodatek za Microsoft Word, ki omogoča prenos razpoznanega govora v tekstovni urejevalnik. Obenem omogoča urejanje dokumenta z uporabo zvočnih ukazov. Ukazi so sestavljeni iz besednih zvez (npr. nova vrstica, briši besedo, itd.), na katere se urejevalnik ustrezno odzove. Sometimes we find that writing by hand or by keyboard can be impractical, tedious, or outright difficult due to one’s physical inability. In such circumstances, it would be beneficial to use a solution which takes advantage of one’s voice for dictation. In this bachelor thesis we covered the development of a dictation application for Slovene and its integration with Microsoft Word. More specifically, we developed an add-in for Microsoft Word, which can transfer speech-recognized text to the text editor. Furthermore, it is capable of editing the text by use of voice commands. The commands consist of phrases (new line, delete word, etc.), to which the application reacts accordingly.

V diplomskem delu je predstavljen obstoječ informacijski sistem za podporo pri obdelavi podatkov taborniškega orentacijskega tekmovanja, njegova zgodovina, njegov podatkovni model, uporabniški vmesnik, poslovna logika in njegove pomanjkljivosti, zaradi katerih je otežena njegova uporaba. Namen je s pomočjo analize obstoječega informacijskega sistema poiskati rešitve, ki odpravijo predstavljene pomanjkljivosti, pripraviti načrt za implementacijo novega informacijskega sistema in s pomočjo implementacije predstaviti rešitve izpostavljenih pomanjkljivosti. Načrt med drugim vsebuje tudi seznam zahtev, načrt arhitekture in podatkovni model. Rezultat diplomskega dela je informacijski sistem z okrnjenimi funkcionalnostmi, ki je osnova za nadaljni razvoj informacijskega sistema, ki bo vseboval vse željene funkcionalnosti. The thesis presents the existing information system for data processing of the orienteering competition, its history, data model, user interface, business logic and shortcomings that make its use difficult. The purpose is to use the analysis of the existing information system to find solutions that solve the presented shortcomings, prepare a plan for the implementation of the new information system and to present the solutions of the exposed shortcomings through the implementation. The plan also includes a list of requirements, an architecture plan and a data model. The result is an information system with reduced functionalities, which is the basis for further development of an information system that will have all the desired functionalities.

Diplomska naloga opisuje postopek razvoja zalednega dela ter mobilne in spletne aplikacije v izbranih tehnologijah in ogrodjih. Veliko ljudi iz skupnosti gluhih in naglušnih ima težave pri opravljanju vsakdanjih opravil kot je npr. udeležba in aktivno sodelovanje na sestanku. Cilj aplikacije je s pomočjo sodobnih spletnih tehnologij ter uporabo in integracijo prodkutov umetne inteligence v uporabniku prijazne aplikacije, olajšati ter digitalizirati sam proces izvedbe in participacije pri sestankih. The thesis describes the development process and technologies used in process of developing backend, mobile and web application. Many people from the deaf and hard of hearing community have difficulties performing everyday tasks such as e.g active participation at meetings. Main goal of this application is to, with help and use of modern web technologies and integration of artificial intelligence products to user friendly apps, facilitate and digitalize process of carrying out and participating at meetings.

V diplomski nalogi se lotili reševanja problema regulacije temperature v prostoru. Opisali smo postopek izdelave pametnega termostata, namenjenega krmiljenju centralnega ogrevanja v bivalnih objektih. Deluje s pomočjo računalnika Raspberry Pi Zero W, ki skrbi za merjenje temperature v prostoru, vklapljanje ogrevanja in opravlja vlogo spletnega strežnika. Nastavljanje delovanja termostata poteka preko spletnega uporabniškega vmesnika. Želeno temperaturo prostora lahko nastavimo ročno ali pa nastavimo tedenski urnik, po katerem bo termostat deloval. Termostat podpira tudi beleženje temperature in napovedovanje stroškov tako, da beleži, koliko časa je bilo ogrevanje vklopljeno. Ti podatki se shranjujejo v podatkovno bazo SQLite, ki se prav tako nahaja na RPi. In this thesis, we have tackled the problem of room temperature control. We described the process of building a smart thermostat for controlling central heating in residential buildings. It works with the help of a Raspberry Pi Zero W computer, which takes care of measuring the room temperature, switching on the heating and acting as a web server. The thermostat is configured via a web-based user interface. The desired room temperature can be set manually or a weekly schedule can be set for the thermostat to run. The thermostat also supports temperature logging and cost forecasting by logging how long the heating has been on. This data is stored in a SQLite database, also located on the RPi.

Integracije več socialnih omrežij znotraj enega produkta so pogoste, poznavanje specifik vseh vključenih API-jev pa precej zamudno. V diplomskem delu naslovimo ta problem z izgradnjo API-ja, ki združuje več API-jev socialnih omrežij. Pogledali smo si vrste in zgodovinski razvoj API-jev, primerjali funkcionalnosti najbolj uporabljanih socialnih omrežij ter raziskali, kakšne rešitve, ki naslavljajo isti problem kot naša, že obstajajo. Ustvarili smo splošni API za dostop do socialnih omrežij, ki združuje API-je socialnih omrežij Facebook, Instagram, LinkedIn in Twitter. API omogoča pridobitev podatkov o profilu prijavljenega uporabnika in njegovih objavah na vseh vključenih platformah, objavljanje na socialnih omrežjih Facebook in LinkedIn ter odziv na obstoječe objave z všečkom ali komentarjem na platformah LinkedIn in Twitter. Integrating multiple social media with a single product is a common practice and knowing the specific of all integrated APIs is quite time consuming. In the theis we address this problem by building an API that combines several social network APIs. We looked at the types and historical development of APIs, compared the functionalities of the most widely used social networks, and explored what solutions addressing the same problem as ours already exist. We have created a general API for accessing social networks that combines the APIs of social networks Facebook, Instagram, LinkedIn and Twitter. The API offers functionality for obtaining information about the registered user's profile and their posts on all included platforms, posting on social networks Facebook and LinkedIn, and responding to existing posts with a like or comment on the platforms LinkedIn and Twitter.

Cilj diplomske naloge je bil razvoj sistema za nadzor in upravljanje pomožnega ogrevanja počitniškega objekta na daljavo. Želeli smo nadomestiti trenutni sistem vzdrževanja temperature prostorov nad lediščem z namenom preprečevanja poškodb vodovodne napeljave objekta zaradi zmrzovanja vode v vodovodnih ceveh. Prav tako smo želeli smo izdelati sistem, ki vsebuje vse dosedanje zmožnosti, je po preprostosti uporabe konkurenčen tistim na tržišču, vendar se za razliko od omenjenih bolje prilagaja potrebam določenega objekta (npr. preprečevanje zmrzali). Po preučitvi področja tehnologije IoT in obstoječih sistemov, smo v načrtu sistema definirali zahteve in se odločili za izdelavo lastnega prototipa sistema, sestavljenega iz odprtokodnih komponent programske opreme, lahko dostopne strojne opreme in lastnega prototipa strojne opreme v obliki termostatskih vtičnic IoT. Rezultat dela je prototip odprtokodnega sistema, ki omogoča vse zmožnosti starega sistema, hkrati pa po zaslugi tehnologije IoT prinaša obilico izboljšav z vidika enostavnosti uporabe, zanesljivosti, varnosti, odpornosti na napake, dodane vrednosti za uporabnika ter omogočanja nadgradnje in razširitve področja uporabe v prihodnosti. The aim of the diploma thesis was to develop a system to control and manage the auxiliary heating of a remote holiday facility. We wanted to replace the current system for maintaining room temperature above freezing to prevent damage to the building's plumbing due to water freezing in the water pipes. We also wanted to create a system that included all of the existing features, was competitive with systems on the market in terms of ease of use, but unlike them, was more customizable to the needs of a particular facility (e.g., freeze protection). After researching the field of IoT technology and existing systems, we defined the requirements in the system plan and decided to create our own prototype system consisting of open-source software components, easily accessible hardware, and our own hardware prototype in the form of IoT thermostatic sockets. The result is an open source prototype that enables all the capabilities of the legacy system, while the IoT technology brings many improvements in ease of use, reliability, security, fault tolerance, added value for the user, and enabling upgrades and expansion in the future.

Tema diplomske naloge zajema postopek razvoja mobilne aplikacije za prepisovanje govora z uporabo razvojnega ogrodja React Native. Glavna problematika trga takih aplikacij je pomanjkanje kvalitetnih aplikacij s podporo za slovenščino, saj izdajatelji mobilnih operacijski sistemov ne vlagajo dovolj virov za nadgrajevanje vgrajenih modelov za razpoznavo govora, ki trenutno podpirajo le ozek nabor jezikov, zato je bil cilj tega dela, da se oblikuje in razvije funkcionalno aplikacijo z intuitivnim uporabniškim vmesnikom, ki skupaj z uporabo obstoječega modela za razpoznavo slovenščine poskuša rešiti omenjen problem. Predstavljeno rešitev je poskusno uporabljala skupina uporabnikov, ki so potem po določenem časovnem obdobju ocenili svojo izkušnjo z aplikacijo. Na podlagi teh informacij smo prišli do ugotovitev o kvaliteti predstavljene rešitve. The topic of this thesis covers the process of developing a mobile application for speech transcription using the React Native development framework. The main problem of the market of such applications is the lack of support for the Slovene language, as publishers of mobile operating systems do not invest enough resources to upgrade embedded speech recognition models, which currently support only a narrow set of languages. The main goal of this paper is to develop a functional application with an intuitive user interface, which together with the use of the existing model for the recognition of Slovene tries to solve the aforementioned problem. The presented solution was used experimentally by a group of users, who then evaluated their experience with the application after a certain period of time. Based on this information, we came to conclusions about the quality of the presented solution.

Klepetalniki in virtualni pomočniki postajajo vedno bolj prisotni v našem življenju. Uporabnikom omogočajo komunikacijo v govorjenem ali pisanem naravnem jeziku prek različnih kanalov za sporočanje. Slovenščine zaradi majhnega števila ljudi, ki jo uporabljamo za komuniciranje, globalno dostopni pametni pomočniki še ne podpirajo. V našem delu smo razvili slovenskega virtualnega pomočnika za upravljanje pametnega doma, ki lahko z uporabo jezikovnega modela za razumevanje naravnega jezika določi namen in entitete v sporočilu uporabnika ter pri tem upošteva kontekst celotnega pogovora. Z uporabo pogovornega modela določi ime akcije, ki je zadolžena za generiranje odgovora. V akcijah smo implementirali različna znanja, s katerimi je lahko uporabnik pridobil številne informacije in izvajal različna opravila. Razvili in evalvirali smo več različnih modelov za klasifikacijo namena in prepoznavo entitet. Najvišjo uspešnost pri klasifikaciji namena smo dosegli z uporabo vektorskih vložitev jezikovnega modela SloBERTa (ocena F1 = 0,900). Pri ekstrakciji entitet smo najvišjo uspešnost dosegli z uporabo vektorskih vložitev modela fastText (ocena F1 = 0,924). Chatbots and virtual assistants are becoming more and more present in our lives. They allow users to communicate in spoken or written natural language, through various communication channels. Slovenian language is poorly supported by globally used smart assistants, due to the small number of people that use it for communication. We have developed a Slovenian virtual assistant for smart home management. The assistant understands natural language and uses a language model to classify the purpose and entities in the user's message by taking the context of the entire conversation into account. It uses a conversational model to determine the name of the action responsible for generating the response. We used actions to implement various skills. They enable users to obtain various information and perform different tasks. In our thesis, we have developed and evaluated several different models for intent classification and entity extraction. The highest performance in the intent classification was achieved by using word embeddings from the SloBERTa language model (F1 score = 0,900). In the extraction of entities, the highest performance was achieved by using word embeddings from the fastText model (F1 score = 0.924).

In our research, we focused on analysing the stock market events and the price movements on market events. We researched the quarterly corporate earning announcement events as well as dividend declaration events and investigated any correlations they might have to the publicly available stock data and stock prices. Furthermore, we analysed how the public sentiment, notably Twitter and Google Trends, can be utilised to help estimate the correlation between event and the stock price movement. To process the data, we utilised the machine learning algorithm logistic regression due to its simplicity and robustness. Our findings were that the earning announcement event outcome could, to some extent, be predicted but the actual movement is not as important as the size of the movement. We also successfully utilised public sentiment to improve our results further. V naši raziskavi smo se osredotočili na analizo delniških trgov in njihova gibanja cen ob dogodkih na trgih. Raziskali smo tako dogodke povezane s četrtletnimi poročili o zaslužku kot napovedi o dividendah ter poglobljeno preučili njihovo korelacijo med javno dostopnimi delniškimi podatki in samo ceno delnic. Analizirali smo tudi kako bi lahko javni sentiment z osredotočenjem na Twitter in Google Trends pomagal izboljšati ocene korelacij med samim dogodkom ter gibanjem cene. Za procesiranje podatkov smo uporabili algoritme strojnega učenja logistična regresija, saj je algoritem zelo preprost ter robusten. V raziskavi smo ugotovili, da je možno delno napovedati gibanje cene delnic po dogodku, vendar je velikost gibanja pomembnejša od same smeri gibanja. Uspešno smo tudi uporabili javni sentiment za nadaljnje izboljšanje rezultatov.

Dandanes se mnoga podjetja v industriji srečujejo s potrebo po digitalni transformaciji. Hiter razvoj na področju interneta stvari prinaša številne tehnologije, ki pripomorejo k zajemu podatkov o dogajanju v proizvodnji in prenos teh v zaledne sisteme, kjer jih lahko povežemo s podatki o poslovanju. Problem povezljivosti naprav se rešuje z uporabo IoT prehodov. Sprva so IoT prehodi služili predvsem za povezovanje naprav danes pa se povečujejo zahteve po dodatni funkcionalnosti, ki naj bi podpirala nameščanje poljubne programske opreme na prehode in s tem omogočala transformacijo navadnega prehoda v prehod z dodano lokalno inteligenco. Razdelali smo arhitekturo takšnega IoT prehoda ter definirali osnovne funkcionalnosti. Na podlagi razdelane arhitekture smo razvili programsko opremo za transformacijo naprav v IoT prehod, oblačno aplikacijo za potrebe registracije in povezave IoT prehodov in omogočili uporabniku uporabo oblačne platforme za potrebe analize in izvoza podatkov. Internet of things had a rapid development over past few years. We can use different technologies for connecting devices and use them for data gathering. This data is usually sent to the cloud systems for analysis and storage. Different devices use different communication protocols therefore device interoperability was always a problem in internet of things. This problem was usually solved with the use of IoT gateways that acted like a bridge for connecting different devices to cloud systems. Today these gateways are required to perform additional tasks and are becoming smarter. We have designed a software architecture for a smart IoT gateway and defined basic functionalities. According to our design we have implemented a prototype of a IoT gateway, cloud application for connecting IoT gateways and overall system overview. Besides that we have enabled users to use full power of cloud platform for data analysis, storage and export.

Razpoznava govora je sistem, ki omogoča avtomatsko pretvorbo govora v besedilo. Izhod takšnega sistema je surovo besedilo brez velikih začetnic, ločil in ostalih oblikovnih lastnosti. Ker je takšno besedilo nepregledno, ročno urejanje pa zahteva veliko dela, so se uveljavile različne metode, ki omenjene težave rešujejo avtomatsko. Takšni sistemi lahko temeljijo na različnih metodah, vendar so se v zadnjem času predvsem zaradi dobrih rezultatov uveljavili različni tipi nevronskih mrež. Tako smo v sklopu magistrskega dela implementirali sistem, ki za svoje delovanje uporablja rekurenčne nevronske mreže. Preizkusili smo ga z različnimi vektorskimi vložitvami, kot so GloVe, ELMO in BERT. Implementirali smo tudi spletno storitev, ki omogoča, da sistem enostavno integriramo v različne storitve, kot je npr. že prej omenjena avtomatska razpoznava govora. Speech recognition is a system that allows for automatic conversion of speech into written text. Such systems typicaly return raw text without any formatting such as capital letters or punctuation symbols. Because such text is unreadable and it also requires a lot of work to edit manually, various methods have been introduced that solve these problems automatically. Such systems can be based on a variety of methods. However, due to good results they provide, different types of neural networks are mainly used nowdays. As part of the master's thesis, we have implemented a system that uses recurrent neural network to predict punctuation symbols in raw unpunctuated text. We have tried it with different word embeddings such as GloVe, ELMO and BERT. We have also implemented a web service that allows us to easily integrate the system into various other services, such as automatic speech recognition.

Zagotavljanje zdravstvene oskrbe je zaradi staranja prebivalstva in posledično višjih stroškov zdravljenja podvrženo čedalje večjim izzivom. Trenutni modeli zdravstvenega zavarovanja v Sloveniji temeljijo na principu medgeneracijske solidarnosti in so vzdržni le v primeru, ko je delež delovno aktivne populacije ustrezno velik. Zato ne čudi dejstvo, da pospešeno iščemo rešitve in nove modele zagotavljanja zdravstvenega varstva. V tem kontekstu se velikokrat omenja telemedicino, oskrbovana stanovanja in oskrbo pacientov na daljavo. Žal se telemedicinske storitve v mnogih evropskih državah uvajajo zelo počasi in Slovenija tu ni izjema. Problem je zakonodaja, ki izvajalcem zdravstvenih storitev trenutno ne priznava stroškov oddaljenega spremljanja pacientov. Cilj diplomske naloge pa ni sprememba zakonodaje, temveč izdelava načrta informacijskega sistema za podporo telemedicinskim storitvam, kar bo, upamo, v pomoč slovenskim razvojnim podjetjem pri pripravi rešitev za to področje v prihodnosti. V ta namen smo najprej opisali že obstoječe sisteme za oskrbo pacientov na daljavo. Sledi njihova primerjava, ki velja za osnovo pri izdelavi načrta informacijskega sistema, nato pa smo predstavili načrt, v katerem smo se informacijskemu sistemu posvetili s procesnega, arhitekturnega, funkcionalnega in varnostnega vidika. As the population is aging, costs of medical services are rising quickly. Current healthcare models in Slovenia are based on intergenerational solidarity and are sustainable only when the share of actively working population is high enough. As a consequence, we are constantly looking for new healthcare models, which would solve this problem. Telemedicine, sheltered housing and remote patient monitoring are often discussed as possible solutions. In Europe, telemedicine is being deployed rather slowly, and Slovenia is no exception. The problem lies in the legislation, which does not provide funding for remote patient monitoring. The goal of the thesis is not to change the legislation, but to develop a plan for an information system which would enable deployment of telemedicine services across the country in the future, when there will be no legal drawbacks. Hopefully, the plan will help development companies with developing solutions for this area of expertise. With this goal in mind, we began with describing exisitng remote patient monitoring systems. We made a comparison of the systems, which served as the base for our plan. Finally, we have thoroughly described the plan, where we focused on the processes, architecture, funcionality and security that the information system will have to provide.

Obravnavan je potek razvoja sodobne aplikacije v oblaku, ki je namenjena piscem spletnih zgodb. Seznanimo se s tehnologijami in arhitekturami, o katerih se dandanes pogosto sliˇsi, ko je govora o računalništvu v oblaku – različni tipi podatkovnih baz (baze SQL, NoSQL, ipd.), programski jeziki (Java, JavaScript – Node.JS, Go), arhitekture aplikacij (monolitna, mikro-storitvena, funkcijska, dogodkovno-vodena). Razvoj zaobjema analizo funkcionalnih in nefunkcionalnih zahtev aplikacije, pripravo načrta arhitekture aplikacije in podatkovnih modelov, izbor ustreznih tehnologij, vzpostavitve razvojnega okolja, pripravo osnovne strukture projekta, razvoj uporabniškega vmesnika in zalednega sistema. Uporabljen izbor je evalviran na podlagi kompleksnosti razvoja, ustreznosti definiranim zahtevnostim in stroškom obratovanja aplikacije. Aplikacija je razvita za platformo oblačnih storitev Google Cloud Platform, uporabniški vmesnik z uporabo ogrodja Angular. Addressed is the course of developing a modern cloud computing application, meant for web-novels writers. We are introduced to modern-day technologies and architectures for cloud computing–different types of databases (traditional SQL, modern NoSQL, etc.), programming languages (Java, JavaScript–Node.JS, Go), application architectures (monolithic, microservices, functional, event-driven). The development comprises of analyzing functional and non-functional requirements, preparing architectural design and data models, selecting appropriate technologies, establishing the development environment, preparing a basic project structure, developing user interface and back-end system. Selection is evaluated based on development complexity, compliance with defined requirements, and application deployment costs. Application is developed for Google Cloud Platform, user interface using Angular framework.

Software development is a complex and creative process. In contrast to a typical business process it tends to be more dynamic and dependent on a number of circumstances. Empirical studies show that companies still don’t document their development practices, or if they do, these are not up-to-date and do not reflect how they really develop software. On the other hand, various supporting tools such as issue tracking system, revision control system, document management system, etc. are used by developers and project managers during their work, capturing a vast body of knowledge about how a software development process has been performed. The main objective of this dissertation is to propose an approach that can help companies in documenting their real development practice. Comparing to existing approaches that require substantial effort on the side of project members, our approach extracts information on development practice directly from software repositories. Five companies have been studied to identify information that can be retrieved from software repositories. Based on this, an approach to reconstruct development practice has been developed. The approach has been evaluated on a real software repository shared by an additional company. The results confirm that software repository information suffice for the reconstruction of various aspects of development process, i.e. disciplines, activities, user roles, and artifacts. Razvoj programske opreme je kompleksen in ustvarjalen proces. V primerjavi s tipičnim poslovnim procesom je bolj dinamičen in odvisen od številnih okoliščin. Empirične študije kažejo, da podjetja še vedno ne dokumentirajo svoje razvojne prakse, če pa že, le-te ne vzdržujejo in posledično ne odražajo, kako dejansko razvijajo programsko opremo. Po drugi strani pa razvijalci in vodje projektov med svojim delom uporabljajo različna podporna orodja, kot so sistem za sledenje zahtevkom, sistem za nadzor verzij, sistem za upravljanje dokumentov, itd., ki zajamejo veliko znanja o tem, kako je bil izveden proces razvoja programske opreme. Glavni cilj pričujoče disertacije je predlagati pristop, ki lahko pomaga podjetjem pri dokumentiranju njihove dejanske razvojne prakse. V primerjavi z obstoječimi pristopi, ki zahtevajo veliko napora na strani članov projekta, naš pristop rekonstruira informacije o razvojni praksi neposredno iz programskih repozitorijev. Na podlagi podatkov petih podjetij so bile identificirane informacije, ki jih je mogoče pridobiti iz programskih repozitorijev. Na podlagi tega je bil razvit pristop za rekonstrukcijo razvojne prakse. Pristop je bil evalviran na resničnem programskem repozitoriju, ki ga je zagotovilo dodatno podjetje. Rezultati potrjujejo, da informacije iz programskih repozitorijev zadostujejo za rekonstrukcijo različnih vidikov razvojnega procesa, tj. disciplin, aktivnosti, uporabniških vlog in artefaktov.

Danes ima uživanje zdrave in uravnotežene prehrane vse večji pomen. Eno izmed ključnih orodji pri doseganju omenjenega cilja je beleženje vrste in količine zaužite hrane. V okviru diplomske naloge najprej predstavimo razvoj klepetalnikov na splošno. Osrednja tema naloge pa je domensko omejeni klepetalnik za govorni vnos zaužite hrane. Osnova za njegovo delovanje je podatkovna baza s skoraj tisoč živili. Z uporabo klepetalnika želimo od uporabnika pridobiti dovolj podatkov, ki nam bodo omogočili identifikacijo živil, ki so bila vključena v posamezen obrok. Pri implementaciji klepetalnika smo uporabili nekaj metod za obdelavo naravnega jezika, kot so lematizacija, kosinusna razdalja, iskanje nizov in Levenshteinova razdalja. Klepetalnik je zmožen sistematično postavljati vprašanja ob nejasnem oz. nepopolnem govornem vnosu. Prototip rešitve je razvit v okolju Java Swing in je na voljo za uporabo na osebnih računalnikih. Today the importance of having a healthy, nutritious diet is greater than ever. One of the main tools in achieving this goal is an efficient way to record the meals one has consumed throughout the day. This diploma thesis gives a brief overview of chatbots in general, but focuses on implementing the logic for a domain specific chatbot. The domain is determined by a database containing almost one thousand foods. The goal is to gather enough information from the user to identify the foods consumed during a particular meal. We use natural language processing (NLP) methods such as lemmatisation, cosine distance, string matching and levenshtein distance. The chatbot is also capable of forming systematical questions when the speech input is incomplete or unclear. The prototype of the chatbot is available as a Java Swing Application for personal computers.

Tehnologije za postavitev pametnih domov so v zadnjih letih postale vse bolj priljubljene in razˇsirjene. Razlog za to je najverjetneje tudi razvoj cenovno dostopnih naprav za pametne domove, ki se poveˇzejo v omreˇzje in tako uporabnikom omogoˇcajo njihov nadzor na daljavo. Namestitev je obiˇcajno uporabnikom prijazna, vendar pa mnogo proizvajalcev za njihovo upravljanje zahteva lastno aplikacijo. Celostne reˇsitve za postavitev pametne hiˇse, ki ˇze od nekdaj omogoˇcajo centralni nadzor, pa so ostale drage. V okviru diplomske naloge smo razvili sistem, ki centralizira nadzor naprav za pametne domove na eno mesto. Slednje se vanj povezujejo preko omreˇzja s pomoˇcjo naˇse knjiˇznice oziroma predhodno definiranega protokola za komunikacijo. Razvili smo tudi vmesnik, ki uporabniku omogoˇca upravljanje pametnega doma. Sistem teˇce na mini raˇcunalniku Raspberry Pi. Nanj se preko omreˇzja povezujejo vgrajene naprave, ki temeljijo na projektu Arduino. Za slednjega smo razvili tudi knjiˇznico za komunikacijo s sistemom In the last few years, we can observe the rise of popularity of smart homes and related technologies. One of the reasons for this is the availability of affordable smart home devices, which connect to the network and thus allow users to control them from anywhere. Installation of these devices is usually very user-friendly and simple, so that people can install them themselves. However, most manufacturers require them to use their app to control these devices. Fully managed systems with a single control point have existed for a long time, but remain expensive and not as widely available. For this thesis, we have developed a system, which centralizes control of the smart home devices to a single interface. These devices connect to the system via network with the help of our library, based on our communication protocol. We have also designed and developed an interface that the user will use for controlling his smart home. The system runs on a small computer called Raspberry Pi and connects to embedded devices, based on Arduino.

Področje računalništva na robu je definirano zelo ohlapno v primerjavi z uveljavljenimi têrmini na področju modernega računalništva v oblaku. Večina poskusov ustvarjanja novih rešitev je osnovana na internetu stvari, ki so nato občasno razširjene z zmožnostjo obdelave podatkov na robu, izven oblaka. Take rešitve so omejene na obdelavo le senzorskih podatkov. Skozi obsežen pregled raziskav v obstoječi literaturi in pregled obstoječih implementacij podobnih rešitev ugotovimo, da nobena ne nudi podpore za infrastrukturo na robu, ki je nižjenivojska kot platforma za razvoj aplikacij. V delu načrtujemo in implementiramo lastno rešitev za upravljanje naprav na robu in skozi njo omogočimo transparentno upravljanje skozi Ansible AWX, ki je uveljavljeno orodje za upravljanje s konfiguracijami naprav. Rešitev, poimenovana mast, nudi postopek za priključitev novih naprav v sistem, vzpostavitev omrežja, varnost, shranjevanje podatkov in usklajevanje časa med napravami v gruči in omogoča enostavno razširitev in priključitev v obstoječe vzorce računalništva v oblaku. Evalvacija pokaže, da, v primerjavi z obstoječimi rešitvami, mast nudi nov in močan pristop k upravljanju z napravami na robu, ne glede na višjenivojsko uporabo. The field of edge computing is very loosely defined, as opposed to the well--established terms used in the modern cloud. Most forays into creating new solutions base themselves in IoT and then occasionally provide the ability to process data at the edge, but are then constrained to processing only data coming from sensors. Through an extensive overview of literature and existing implementations, we find that no current solution encompasses the area of edge infrastructures, which are lower--level than application platforms. We design and implement our own solution for managing edge devices and enable transparent management through Ansible AWX, an existing tool for machine configuration management. The implementation, dubbed mast, provides device onboarding, networking, security, storage and time synchronisation to devices in the cluster, and allows a straightforward extension of the cloud paradigm. The evaluation shows that, compared to existing platforms, mast provides a novel and powerful method of managing edge devices, regardless of the workload running on them.

Povezljivost in enostavnost integracije naprav IoT sta ena izmed poglavitnih razlogov, zakaj je v zadnjih letih močno naraslo zanimanje za pametne domove. Starejše tehnologije, kot je KNX, imajo veliko težavo slediti razvoju in zagotavljati enako uporabniško izkušnjo, kot jo danes ponujajo največji proizvajalci na trgu naprav IoT za pametni dom. Veliko prepreko na trgu pametnih domov KNX predstavlja predvsem konfiguracija, saj ta ne vsebuje informacij o končnih napravah v uporabnikovem domu, ampak zgolj konfiguracijo posameznih funkcij. V magistrskem delu smo razvili komponento, ki iz projektne datoteke, kjer je zapisana celotna konfiguracija funkcij v KNX-domu, rekonstruira končne naprave v primerni obliki za vizualizacijo ali glasovno opravljanje naprav v domu. V prvem koraku smo sestavili podatkovno množico, kjer so funkcije razdeljene v 65 različnih razredov funkcionalnosti. Iz krajših tekstovnih razlag v angleškem in nemškem jeziku smo nato z metodami procesiranja naravnega jezika in večrazredno klasifikacijo izdelali napovedni model za predikcijo funkcijskih razredov. Na podlagi napovedanih razredov in preostalih parametrov konfiguracije smo potem rekonstruirali naprave v domu. Na koncu smo razvili aplikacijski modul, ki uporabniku omogoča, da naloži projektno datoteko doma, preko le-te pa je nato zgenerirana ustrezna konfiguracija. Implementirano rešitev smo primerjali z alternativno rešitvijo in ugotovili, da smo z našo metodo precej povečali število pravilno zaznanih naprav. Connectivity and integration simplicity are two of the main features of the IoT market which drive the rising popularity of Smart Home in the last years. Older technologies like KNX are struggling to follow the rapid development and to deliver the same user experience as IoT giants do. The biggest barrier for the KNX market lies in its configuration specifics, which is not aware of end-customer devices included in the home but only saves the configuration of each functionality. In this master's thesis we developed a component, which reconstructs the complete home configuration into a form ready for visualization or voice control of specific devices based on KNX project file, which includes a complete configuration of KNX Smart Home. In the first phase we prepared a dataset of KNX functions, which are distributed into 65 classes. Using natural language processing techniques and multi-class classification algorithms, we then constructed a prediction model for predicting specific function class based on English and German short text function descriptions. Using this information and other parameters from KNX project file we then group functions into meaningful devices included in the home. At the end, we developed an application module, which includes an element for uploading KNX project file based on which we then generate the adequate configuration. Comparison between the implemented module and the alternative solution showed that we have successfully increased the number of correctly detected devices.

Ljudje že stoletja ogrevamo svoja bivališča na najrazličnejše načine. Sodobnejši in človeku prijaznejši načini so na voljo le s pritiskom na gumb in nastavitvijo željene temparature. Večkrat se bivališča grejejo tudi, ko nas ni doma in na temperature, ki nam ne predstavljajo ugodnega počutja. V magistrskem delu predstavljamo nadgradnjo obstoječega sistema toplotne črpalke, ki spremlja naše gibanje tudi izven naših bivališč, izklaplja naše toplotne sisteme, ko zazna, da bomo odsotni dlje časa in ogreje bivališče na nam ugodno temparaturo točno ob prihodu domov. Tako nam o predhodnih ročnih regulacijah ni potrebno popolnoma nič razmišljati. Poleg tega ohranimo precej nepotrebno potrošene energije in s tem posredno varujemo okolje. Ogrevalni sistem smo nadgradili z avtomatiziranim sistemom, ki je povezan z GPS sistemom mobilne naprave in omogoča strojno učenje glede na predhodne situacije. Rezultati analize delovanja novega sistema so pri povprečnih dnevnih zunanjih temperaturah med 7 in 9 stopinj Celzija pokazali tudi več kot 40 % prihrankov električne energije. For centuries, people have been using different types of heating systems. Modern and user-friendly systems are regulated just by pressing a button and setting the right temperature. Our homes can be heated when we are away and often reach unpleasant temperatures. In our master thesis, we presented an upgrade of the existing heat pump system that tracks our movement even when we are away from home. It recognises when we will be away for a longer period of time, switches off our heating as needed, and sets the right temperature at the exact time of our return. That means we do not have to worry about manual controls. In addition, we are protecting the environment and saving energy. The heating system has been upgraded with an automatic system that is connected to a mobile device through GPS and enables machine learning based on past behaviour. Analyses showed that using the new heating system at average outdoor temperatures between 7 and 9 degrees Celsius can reduce energy consumption for more than 40 %.

Diplomsko delo se ukvarja z razvojem prototipnega sistema za glasovno upravljanje aplikacije na android napravi. Na spletu se pojavljajo večinoma plačljive storitve razpoznave govora (tudi za slovenski jezik), ki jih ponujajo tehnološki giganti in delujejo večinoma preko interneta. Odprtokodni sistemi za razpoznavo govora so slabo dokumentirani in vsebujejo podporo zgolj za svetovne jezike, kot sta na primer angleščina in nemščina. Zato je bil izveden in opisan celoten postopek razvoja razpoznave govora na android napravi za slovenski jezik na podlagi odprtokodnega orodja CMU Sphinx, ki deluje brez internetne povezave. Z uporabo orodij CMU Sphinx je bil razvit akustični model za omejen nabor slovenskih ukazov na podlagi enega govorca. Ta akustični model je bil nato integriran v preprosto demonstracijsko android aplikacijo, kjer je bilo s pomočjo knjižnice PocketSphinx implementirano razpoznavanje ukazov v slovenskem jeziku. Rezultati testiranj so bili izredno uspešni in so pokazali hitro ter natančno delovanje razpoznave govornih ukazov. The diploma thesis is focused on the development of a prototype system for voice control of an application on an android device. On internet majority of services for speech recognition offered by technological giants are payable and mostly works with usage of internet. Open source speech recognition systems are not well documented and support only world languages like english or german. Therefore, the whole process of speech recognition development on android device for the Slovenian language was created and described on the basis of open source tools CMU Sphinx, which are operating without an internet connection, which has not yet been addressed. Using the CMU Sphinx tools, an acoustic model was developed for a limited set of Slovene commands based on one speaker. This acoustic model was then integrated into a simple demonstration android application where the recognition of commands in the Slovene language was implemented using the PocketSphinx library. Test results were extremely successful and showed the fast and accurate operation of voice command recognition.

Vedno več ljudi se zaveda pomembnosti zdrave in raznolike prehrane v njihovem vsakdanjem življenju. Pogosto želijo imeti pregled nad tem, kaj so pojedli, saj tako lažje spremljajo prehranske navade. Velika večina ljudi ima vedno s seboj mobilni telefon, s katerim lahko fotografirajo jed, zato smo se odločil za izdelavo mobilne aplikacije, ki bi uporabnikom omogočala enostaven zajem zaužite hrane. V diplomskem delu smo najprej predstavili podobne aplikacije in njihove glavne funkcionalnosti. Nato smo opisali tehnologije in programske jezike, ki smo jih uporabili v okviru diplomskega dela. Predstavili smo žični načrt aplikacije in izpostavili pomen povezovanja mobilnih aplikacij z zunanjimi sistemi. Nato smo prešli na arhitekturo, kjer smo opisali dva zaledna sistema, ki ju uporabljamo -- VitaBits API in FoodDiary API. Nadaljevali smo s predstavitvijo videza in funkcionalnosti razvite aplikacije. Opisali smo glavne zaslonske maske in njihov pomen. Zaključili smo s pregledom opravljenega dela in predstavili uporabnost aplikacije. More and more people are becoming aware of the importance of a healthy and varied diet in their daily lives. They often want to have an overview of what they have eaten as it makes it easier to follow their eating habits. The vast majority of people always have a mobile phone with them, which allows them to take pictures of a dish, so we decided to create a mobile application which would allow users to easily capture the food that they ate. In this diploma thesis, we first presented similar applications and their main functionalities. We then described the technologies and programming languages used in the diploma thesis. We introduced the application wireframe and highlighted the importance of connecting mobile applications to external systems. We then moved on to the architecture where we described two back end systems that we use, the VitaBits API and the FoodDiary API. We continued with an introduction to the appearance and functionality of the developed application. We described the main application pages and their importance. We concluded with an overview of the work that we have done and outlined the utility of the application.

Poslovno domeno pričujočega diplomskega dela predstavlja sistem souporabe 100 % električnih vozil, ki predstavlja 24/7 samopostrežno možnost najema vozil preko aplikacije na pametnem telefonu. Pri spremljanju razvoja sistema sem identificiral naslednji problem: z večanjem števila uporabnikov, vključenih vozil in prevzemno-vračilnih lokacij ter funkcionalnosti se veča tudi njegova kompleksnost, pri čemer morajo upravljalci storitve na različnih ravneh z namenom učinkovitega upravljanja ohraniti nadzor in vpogled v ključne parametre delovanja sistema. Cilj diplomskega dela je predstavitev rešitve, nadzorne plošče, ki podjetju in upravljalcem omogoča natančen vpogled v delovanje sistema ter na pregleden in enostaven način prikazuje ključne parametre sistema glede na raven in potrebe upravljalca, s čimer je omogočena tudi hitrejša odzivnost in zaznavanje morebitnih odstopanj. Pri doseganju cilja sem uporabil sodobne tehnologije za razvoj spletnih aplikacij ter uveljavljene komunikacijske protokole. Rešitev sem tudi analiziral in predstavil možnosti za prihodnje nadgradnje na tehničnem nivoju in na nivoju poslovne domene. The business domain of this final paper presents a sharing system of 100% electric vehicles, which offers a 24/7 self-service vehicle rental through a smart phone application. In monitoring the system development, I identified the following problems: by increasing the number of users, included vehicles, pick-up and drop-off locations, and its functionality, the system’s complexity also increases, which means the service managers have to maintain the control and insight into key parameters of the system operation in order to manage it effectively. The goal of this final paper is to present a solution—a control panel that provides the company and system managers with an accurate insight into the system operation, and a transparent and simple display of the system’s key parameters according to the user’s level and requirements, which also enables a faster responsiveness and detects potential deviations. In achieving the goal, I used modern web application development technologies and established communication protocols. I analysed the solution and presented options for future upgrades on the technical and business domain level.

Za izdelavo diplomskega dela sem se odločil, ker sem prostovoljni gasilec v domačem društvu. Trenutno stanje na tem področju je slabo. Vsako leto opravljamo preventivni pregled vseh hidrantov in vodnih rezervoarjev na našem področju in ugotovitve zapisujemo na papir. Ker so ti papirji pospravljeni v arhivu, nam v primeru intervencije ne pomagajo dosti. Po drugi strani pa ima vsak gasilec takrat pri sebi telefon, na katerem bi lahko za katero koli lokacijo v občini pridobil najbližje hidrante in stanje teh. Pri uporabi telefona skupaj z oznako NFC, pa sem se odloči v sklopu aplikacije izdelati časovnik, ki ga nastavimo, ko grejo gasilci v notranji napad in nas opozori po določenem času. I decided to write this diploma thesis because I am a volunteer firefighter in my local fire department. Each year we conduct preventive check-ups of hydrants and water tanks in our region. We write down our observations on a piece of paper. Since these documents are later stored in an archive, they are not very helpful during an intervention. However each firefighter owns a mobile phone that could potentially be used to access the information on the nearest hydrants and their condition, for any location within the municipality. For phones with NFC, I have decided to make a timer as part of the applicationl. The timer can be set when firefighters enter a burning building, and it will give off a warning after a certain period of time has passed.