In this thesis, the performance of a number of models used to predict the result in soccer games has been investigated, using data on the previously played games in the league. Established models were implemented, and tested on a wide set of soccer leagues during several years. The performance of each model was measured using the likelihood ratio against a simple baseline distribution. The performance of these models was then analyzed to find systematic differences correlating with some properties of a soccer league, such as average number of goals in the league, and determine which models overall performed best. The results showed that such differences do exist, correlating with the average number of goals in a league as well as the variance in performance among teams in the league. Additionally, statistically significant differences in the performance of some models were established. / I denna rapport undersöktes prestandan hos ett antal modeller för att förutsäga resultat i fotbollsmatcher, tränade på resultaten i tidigare matcher i ligan. Etablerade modeller implementerades och testades sedan på flera årgångar av ett brett urval av fotbollsligor. Prestandan för varje modell mättes som en likelihood-ratio mot en enkel basdistribution. Modellernas prestanda analyserades sedan för att hitta systematiska skillnader som korrelerar med någon viss egenskap hos en fotbollsliga, t.ex. genomsnittligt antal mål i ligan, samt att avgöra vilka modeller som presterade bäst. Resultaten visade att sådana skillnader finns, och att prestandan korrelerar med dels genomsnittligt antal mål i ligan, men även variansen hos prestandan för lagen i ligan. Dessutom hittades statistiskt signifikanta skillnader i prestanda mellan några av modellerna.

Message queues are data structures that can be distributed over a network to send messages between different end points. These message queues are used for example in micro service based architectures for communication. Some message queues sends one message at a time and some queues batch messages together and sends them in a chunk. This study focused on batch file based message queues and how compression affects latency and throughput versus no compression. The use of the right compression method is vital for reducing network traffic, CPU and memory usage. This study performed tests where data of different sizes were sent from a producer to the queue and then consumed by a consumer. These three components were placed on a locked local machine. Tests were done with different number of messages sent and different sizes of each message. The data that were sent was a map of the Faroe Islands represented as a XML file to create fair compression ratios. The study concluded that at no point had no compression lower latency or higher throughput compared to a tested compression method.

The development and wide spread of Internet browsers and technologies make them a tool that can be used for many scientific problems. This raises the question of whether Internet browsers, together with WebGPU and WebRTC, can be used to do scalable computing in a distributed cluster. This thesis answers the question by implementing a peer-to-peer cluster and testing it with two problems, Matrix multiplication and Mandelbrot sets generation. The experimental results show that computing embarrassingly parallel problems are scalable with more than 75% efficiency.

Sentiment classification is generally used for many purposes such as business related aims and opinion gathering. In overall, since most text sources in the world wide web were written in English, available sentiments classifiers were trained on datasets written in English but rarely in other languages. This raised a curiosity and interest in investigating Sentiment Classification methods to implement on Swedish data. Therefor, this bachelor thesis examined to what extent the connotation of Swedish sentiments would be maintained/retained when translated into English. The research question was investigated by comparing the results given by applying Sentiment Classifications techniques. Further, an investigation of the outcomes of a combination of a lexicon based approach and a machine learning based approach by using machine translation on Swedish Tweets was made. The source data was in Swedish and gathered from Twitter, a naive lexicon based approach was used to score the polarity of the Tweets word by word and then a sum of polaritie was calculated.The swedish source data was translated into English, it was run through a supervised machine learning based classifier to where it was scored. In short, the outcomes of this investigation have shown promising results e.g. the translation did not affect the sentiments in a text but rather other circumstances did. These other circumstances was mostly due to cross-lingual sentiment classification problems and supervised machine learning classifiers character.

Reinforcement learning attempts to mimic how humans react to their surrounding environment by giving feedback to software agents based on the actions they take. To test the capabilities of these agents, researches have long regarded board games as a powerful tool. This thesis compares two approaches to reinforcement learning in the board game backgammon, a Q-table and a deep reinforcement network. It was determined which approach surpassed the other in terms of accuracy and convergence rate towards the perceived optimal strategy. The evaluation is performed by training the agents using the self-learning approach. After variable amounts of training sessions, the agents are benchmarked against each other and a third, random agent. The results derived from the study indicate that the convergence rate of the deep learning agent is far superior to that of the Q-table agent. However, the results also indicate that the accuracy of Q-tables is greater than that of deep learning once the former has mapped the environment.

Little research has been done on antonymic relations, a great deal of this has been done by linguists Paradis et al. Gamification was used in natural language acquisition by Bos and Nissim in their 2015 study about noun-noun compound relations, but gamification of information retrieval remains a relatively new field of study. This thesis reproduced work done by Paradis et al. in an attempt to answer the following questions for Swedish antonyms: will reversing word order in antonymic relations affect the strength of said word pair? Will the perceived strength of canonical antonyms have a lower variance than that of non-canonical antonyms? It will also examine whether giving points depending on the agreement with other users reduce the occurrence of extreme points on an ordinal scale? Two parallel studies were conducted, one using a web app which implemented consensus driven rewards, and another utilising a questionnaire. Reversing the order of the words did not to alter the perceived strength of the antonymic pair, which is consistent with the results acquired by Paradis et al. in 2009. Results regarding variance of canonical and non-canonical antonym pairs were inconclusive. An implementation with consensus driven rewards yielded more extreme values than the questionnaire. More research is suggested to improve the strength of the results. / Lite forskning har gjorts om antonymer inom natural language acquisition, och mycket av den forskning som finns om antonymer är gjord av Paradis et al. inom lingvistik. Gamifiering har använts inom natural language acquisition, bland annat av Bos och Nissim i deras studie om relationer hos sammansatta substantiv från 2015. Denna rapport försöker besvara följande frågor om svenska antonymer: spelar ordningen på ord i ett antonympar någon roll i hur parets antonymiska styrka uppfattas? Kommer den uppfattade styrkan hos kanoniska antonympar ha lägre varians än deras ickekanoniska motsvarigheter? Rapporten undersöker även huruvida konsensusdriven poängsättning påverkar förekomsten av extremvärden på en ordinalskala. Två parallella delstudier utfördes, en webapp som implementerade konsensusdriven poängsättning, samt ett frågeformulär utan poängsättning. Ordningen av orden i ett antonympar hade ingen signifikant påverkan på dess uppfattade styrka, i enighet med Paradis et al.s studie från 2009. Resultaten angående kanoniska antonymers varians var inte entydiga. Implementationen med konsensusbaserad poängsättning gav fler extremvärden än frågeformuläret. Eftersom detta var en liten studie behövs vidare undersökning för att stärka resultaten.

Ever since humans started competing with each other, people have tried to accurately predict the outcome of such events. Football is no exception to this and is extra interesting as subject for a project like this with the ever growing amount of data gathered from matches these days. Previously predictors had to make there predictions using there own knowledge and small amounts of data. This report will use this growing amount of data and find out if it is possible to accurately predict the outcome of a football match using the k-Nearest Neighbor algorithm and Linear regression. The algorithms are compared on how accurately they predict the winner of a match, how precise they predict how many goals each team will score and the accuracy of the predicted goal difference. The results are graphed and presented in tables. A discussion analyzes the results and draw the conclusion that booth algorithms could be useful if used with a good model, and that Linear Regression out performs k-NN. / Ända sedan vi människor började tävla mot varandra, har folk försökt förutspå vinnaren i tävlingarna. Fotboll är inget undantag till detta och är extra intressant för den här studien då den tillgängliga mängden data från fotbollsmatcher ständigt ökar. Tidigare har egna kunskaper och små mängder data använts för att förutspå resultaten. Den här rapporten kommer dra nytta av den växande mängden data för att ta reda på om det är möjligt att med hjälp av k-Nearest Neighbor algoritmen och Linjär regression förutspå resultat i fotbollsmatcher. Algoritmerna kommer jämföras utifrån hur exakt de förutspår vinnaren i matcher, hur många mål de båda lagen gör samt hur precist algoritmerna förutspår målskilnaden i matcherna.    Resultaten presenteras både i grafer och i tabeller. En diskusion förs för att analysera resultaten och kommer fram till att båda algoritmerna kan vara användbara om modelen är välkonstruerad, och att Linjär regression är bättre lämpad än k-NN.

Predicting user preferences is a common problem for many companies and services. With the growth of Internet services it becomes both more important and more lucrative being able to predict what products a user would like and then recommend these to them. There are many ways of attempting this, but this study attempts to use random indexing to solve the same problem. Random indexing is a method that has been used successfully when studying the similarity between words, and allows entities to be represented as vectors with relatively small dimensionality. This would allow for fast and memory-efficient implementations of prediction systems. This study uses the dataset Amazon Fine Food Reviews, which contains reviews of products with a rating. It is attempted to predict these ratings, and the result of random indexing is compared to the results on the same dataset using collaborative filtering. Various parameters used in the random indexing method are also varied, to study their effect on the results. These methods are evaluated based on root mean square error and mean absolute error. The results indicate that random indexing does not generate as good results as collaborative filtering. However, the difference is small enough to warrant further study into the other strengths of random indexing, such as speed and memory efficiency. It is theorized that the sparsity of the dataset might have caused the differences in errors between the methods, and with a dense dataset the results might be better.

The time required to find the optimal solution to a problem increases exponentially as thesize and amount of parameters increases. Evolutionary algorithms tackle this problemheuristically by generating better solutions over time. When there is more than oneobjective in a problem, algorithms must generate multiple solutions to fit any preference inspecific objectives. As the amount of objectives increases, the effort required to generategood sets of solutions increases.This study investigated how increasing the amount of objectives impacted fourmulti-objective evolutionary algorithms differently. The algorithms were tested againsttwo different sets of problems with each problem being tested in twenty seven differentcircumstances. The results of these tests were summarized into two different statisticsbased on ranking used to determine if there was any performance changes.The results indicate that some multi-objective evolutionary algorithms havebetter performance against problems with more objectives. The underlying cause andmagnitude in performance difference was not identified. / När storleken och antalet parametrar växer för ett problem växer tidens som krävs för att hitta den optimala lösningen exponentiellt. Evolutionära algoritmer löser detta problem med heuristik genom att generera bättre lösningar iterativt. När problemen har mer än ett mål måste algoritmerna generera flera lösningar för att passa olika preferenser i specifika mål. Mängden arbete som krävs för att generera bra lösningsmängder växer när antalet mål växer för dessa problem. Denna studie undersökte om ökningen av antalet mål påverkade fyra olika multiobjektiva evolutionära algoritmer olika. Algoritmerna testades mot två olika mängder problem varav varje problem testades under tjugosju olika inställningar. För dessa tester sammanfattades resultat i två olika mätningar baserade algoritmernas rangordning i ett antal mätningar för att komma fram till om det var några skillnader i prestanda. Resultaten påpekade att vissa multiobjektiva evolutionära algoritmer har bättre prestanda hos problem med fler mål. Den underliggande anledningen och storleken på prestandaskillnaden kartlagdes inte.

Recommender systems help shape the way the internet is used by leading users directly to the content which will interest them most. Traditionally, collaborative recommender systems based purely on user ratings have been proven to be effective. This report focuses specifically on film recommender systems. It investigates how the film content parameters Actor, Director and Genre can be used to further enhance the accuracy of predictions made by a purely collaborative approach, specifically with regards to the set of films chosen when performing the prediction calculations. The initial results showed that relying solely on content in this selection led to poorer predictions due to a lack of ratings. However, the investigation finds that using a hybrid approach between the two selection techniques with a bias for content solved this problem as well as increasing the overall prediction accuracy by over 11%.