This thesis proposes the design, development and evaluation of a hybrid video recommendation system. The proposed hybrid video recommendation system is based on a graph algorithm called Adsorption. Adsorption is a collaborative filtering algorithm in which relations between users are used to make recommendations. Adsorption is used to generate the base recommendation list. In order to overcome the problems that occur in pure collaborative system, content based filtering is injected. Content based filtering uses the idea of suggesting similar items that matches user preferences. In order to use content based filtering, first, the base recommendation list is updated by removing weak recommendations. Following this, item similarities of the remaining list are calculated and new items are inserted to form the final recommendations. Thus, collaborative recommendations are empowered considering item similarities. Therefore, the developed hybrid system combines both collaborative and content based approaches to produce more effective suggestions.

The main goal of this work is to facilitate the exchange of information among the members of the social network by extending the functionality of the social network. The network should have the functionality that could provide not only the most relevant information to the particular member of the network but also with some additional information that could be of some interest to that member (user of the system). Such functionality can be achieved by extending social network with recommendation management subsystem. In this work, recommendation is defined as an informal or formal statement defining what information should be provided to the particular user. This work covers the method of the formation of personalized recommendations. The formation of them is based on leverage coefficients. This way of formation of recommendations allows defining the most suitable level of flexibility and personalization. This provides selection and presentation of proper additional information. The algorithm of formation of recommendations, described is this work, formally defines the composition of recommendation. Recommendation is saved as substantive element of the social network. It consist of finite set of atomic elements, this feature lets to analyze or modify recommendations, avoiding flexibility and personalization problems, which emerges in others recommendation systems.

More than 50% of the cultivated areas in the Southern Brazil and more than 60% of the cultivable area in Paraguay adopted the no-tillage system. This system went through several improvements, but there are still doubts about the fertilizing recommendation. Nitrogen (N) and potassium (P) are the nutrients absorbed in largest quantities and their deficiencies affect crop productivity. In this sense, the purpose of this study was to develop a nitrogen and potassium fertilizer recommendation system suitable for wheat, corn and soybean under no-tillage system in Paraguay. Seven field experiments were conducted in four Paraguayan Departments, which include the main grain production regions. Soils in these regions are mainly Oxisols , Inceptisols and Ultisols. Experimental design was random blocks with three replications. Treatments consisted of five rates of N for wheat (0, 30, 60, 90 e 120 kg ha-1 of N), five for corn (0, 60, 120, 180 e 240 kg ha-1 of N), and five rates of K (0, 25, 50, 75 e 100 kg ha-1 of K2O) for corn, wheat and soybean. Those nutrients there are not included in the study were properly supplied. For N, curves of these crops were made and maximum technical and economical efficiencies were calculated, as well as lavels of fertilizers recommended for corn and wheat under notillage system. For K, the soil critical level was determined, correlating level of K in the soil (determined by the method Mehlich 1) and the plant response (relative productivity). It was also determined lavels of K needed to attain the critical le vel in the soil and also to maintain it after it was reached. For wheat, it is necessary 35 kg ha-1 of N to reach 3100 kg ha-1 of wheat after soybean, and after corn it is necessary 30 kg ha-1 of N for a productivity of 2100 kg ha-1 of wheat. Wheat responds to residual N applied in corn. For a productivity of 6000 kg ha-1 of corn it is necessary 90 kg ha-1 after the wheat with little straw leftover. On the other hand, when there is a large quantity of straw and the no-tillage system is not stabilized, 130 kg ha-1 of N is needed. For a productivity of 8000 kg ha-1 of corn, 120 kg ha-1 of N is necessary in a consolidated no-tillage system. The critical level of K in the soil as determined by the Mehlich 1 method is 74 mg dm-3 for a relative productivity of 90%. To increase one mg dm-3 the level of K in the soil, 5 kg ha-1 of K2O is needed. At levels above 150 mg dm-3 plants do not respond to K application and only starting rates are recommended for wheat, corn and soybean crops. / O sistema plantio direto é adotado em mais de 50% das áreas do sul do Brasil e em mais de 60% de toda área cultivável do Paraguai. O sistema passou por várias adaptações, mas ainda permanecem dúvidas quanto à recomendação de adubação. O nitrogênio (N) e o potássio (K) são os nutrientes absorvidos em maior quantidade pelas plantas e suas deficiências afetam as produtividades. Neste sentido, o objetivo deste trabalho foi elaborar um sistema de recomendação de adubação nitrogenada e potássica adequado para trigo, milho e soja sob sistema plantio direto no Paraguai. Foram conduzidos sete experimentos em rede, em quatro departamentos do Paraguai, englobando as principais regiões produtoras de grãos. Os solos dessas regiões são principalmente Oxisols, Inceptisols e Ultisols. O esquema experimental foi de blocos ao acaso com três repetições. Os tratamentos constaram de cinco doses de N para trigo (0, 30, 60, 90 e 120 kg ha-1 de N), cinco para milho (0, 60, 120, 180 e 240 kg ha-1 de N) e cinco doses de K (0, 25, 50, 75 e 100 kg ha-1 de K2O) para trigo, milho e soja. Os elementos que não estavam em estudo foram supridos adequadamente. Para N foram determinadas as produtividades das culturas, elaboradas as curvas de resposta e calculadas máxima eficiência técnica e econômica, assim como as doses recomendadas para o milho e o trigo sob sistema plantio direto. Para o K, foi determinado o teor crítico no solo, correlacionando o teor de K no solo (determinado pelo método de análise Mehlich 1) e a resposta das plantas (rendimento relativo). Foram determinadas as doses de K necessárias para alcançar o teor crítico no solo e para mantê -lo depois de atingido. Para o trigo são necessários 35 kg ha-1 de N para atingir 3100 kg ha-1 de trigo após a soja, e após o milho são necessários 30 kg ha-1 de N para produtividade de 2100 kg ha-1 de trigo. O trigo responde ao residual de N aplicado no milho. Para produtividade de 6000 kg ha-1 de milho são necessários 90 kg ha-1 de N após o trigo com pouca palha. Já quando há grande quantidade de palha e o sistema plantio direto não está estabilizado, são necessários 130 kg ha-1 de N. Para produtividades de 8000 kg ha-1 de milho são necessários 120 kg ha-1 de N num sistema plantio direto consolidado. O teor crítico de K no solo determinado pelo método Mehlich 1 é de 74 mg dm-3 para rendimento relativo de 90%. São necessários 5 kg ha-1 de K2O para elevar em um mg dm-3 o teor de K no solo. Com teores acima de 150 mg dm-3 no solo, as plantas geralmente não apresentam resposta à aplicação de K, recomendando-se somente dose de arranque para as culturas de trigo, milho e soja.

Internet Protocol television (IPTV) is a way of delivering television over the Internet, which enables two-way communication between an operator and its users. By using IPTV, users have freedom to choose what content they want to consume and when they want to consume it. For example, users are able to watch TV shows after they have been aired on TV, and they can access content that is not part of any linear TV broadcasts, e.g. movies that are available to rent. This means that, by using IPTV, users can get access to more video content than is possible with the traditional TV distribution formats. However, having more options also means that deciding what to watch becomes more difficult, and it is important that IPTV providers facilitate the process of finding interesting content so that the users find value in using their services. In this thesis, the author investigated how a user’s online social network can be used as a basis for facilitating the discovery of interesting movies in an IPTV environment. The study consisted of two parts, a theoretical and a practical. In the theoretical part, a literature study was carried out in order to obtain knowledge about different recommender system strategies. In addition to the literature study, a number of online social network platforms were identified and empirically studied in order to gain knowledge about what data is possible to gather from them, and how the data can be gathered. In the practical part, a prototype content discovery system, which made use of the gathered data, was designed and built. This was done in order to uncover difficulties that exist with implementing such a system. The study shows that, while it is is possible to gather data from different online social networks, not all of them offer data in a form that is easy to make use of in a content discovery system. Out of the investigated online social networks, Facebook was found to offer data that is the easiest to gather and make use of. The biggest obstacle, from a technical point of view, was found to be the matching of movie titles gathered from the online social network with the movie titles in the database of the IPTV service provider; one reason for this is that movies can have titles in different languages. / Internet Protocol television (IPTV) är ett sätt att leverera tv via Internet, vilket möjliggör tvåvägskommunikation mellan en operatör och dess användare. Genom att använda IPTV har användare friheten att välja vilket innehåll de vill konsumera och när de vill konsumera det. Användare har t.ex. möjlighet att titta på tv program efter att de har sänts på tv, och de kan komma åt innehåll som inte är en del av någon linjär tv-sändning, t.ex. filmer som är tillgängliga att hyra. Detta betyder att användare, genom att använda IPTV, kan få tillgång till mer videoinnhåll än vad som är möjligt med traditionella tv-distributionsformat. Att ha fler valmöjligheter innebär dock även att det blir svårare att bestämma sig för vad man ska titta på, och det är viktigt att IPTV-leverantörer underlättar processen att hitta intressant innehåll så att användarna finner värde i att använda deras tjänster. I detta exjobb undersökte författaren hur en användares sociala nätverk på Internet kan användas som grund för att underlätta upptäckandet av intressanta filmer i en IPTV miljö. Undersökningen bestod av två delar, en teoretisk och en praktisk. I den teoretiska delen genomfördes en litteraturstudie för att få kunskap om olika rekommendationssystemsstrategier. Utöver litteraturstudien identifierades ett antal sociala nätverk på Internet som studerades empiriskt för att få kunskap om vilken data som är möjlig att hämta in från dem och hur datan kan inhämtas. I den praktiska delen utformades och byggdes en prototyp av ett s.k. content discovery system (“system för att upptäcka innehåll”), som använde sig av den insamlade datan. Detta gjordes för att exponera svårigheter som finns med att implementera ett sådant system. Studien visar att, även om det är möjligt att samla in data från olika sociala nätverk på Internet så erbjuder inte alla data i en form som är lätt att använda i ett content discovery system. Av de undersökta sociala nätverkstjänsterna visade det sig att Facebook erbjuder data som är lättast att samla in och använda. Det största hindret, ur ett tekniskt perspektiv, visade sig vara matchningen av filmtitlar som inhämtats från den sociala nätverkstjänsten med filmtitlarna i IPTV-leverantörens databas; en anledning till detta är att filmer kan ha titlar på olika språk.

Sistemas de Recomendação (SR) são softwares capazes de sugerir itens aos usuários com base no histórico de interações de usuários ou por meio de métricas de similaridade que podem ser comparadas por item, usuário ou ambos. Existem diferentes tipos de SR e dentre os que despertam maior interesse deste trabalho estão: SR baseados em conteúdo; SR baseados em conhecimento; e SR baseado em filtro colaborativo. Alcançar resultados adequados às expectativas dos usuários não é uma meta simples devido à subjetividade inerente ao comportamento humano, para isso, SR precisam de soluções eficientes e eficazes para: modelagem dos dados que suportarão a recomendação; recuperação da informação que descrevem os dados; combinação dessas informações dentro de métricas de similaridade, popularidade ou adequabilidade; criação de modelos descritivos dos itens sob recomendação; e evolução da inteligência do sistema de forma que ele seja capaz de aprender a partir da interação com o usuário. A tomada de decisão por um sistema de recomendação é uma tarefa complexa que pode ser implementada a partir da visão de áreas como inteligência artificial e mineração de dados. Dentro da área de inteligência artificial há estudos referentes ao método de raciocínio baseado em casos e da recomendação baseada em casos. No que diz respeito à área de mineração de dados, os SR podem ser construídos a partir de modelos descritivos e realizar tratamento de dados textuais, constituindo formas de criar elementos para compor uma recomendação. Uma forma de minimizar os pontos fracos de uma abordagem, é a adoção de aspectos baseados em uma abordagem híbrida, que neste trabalho considera-se: tirar proveito dos diferentes tipos de SR; usar técnicas de resolução de problemas; e combinar recursos provenientes das diferentes fontes para compor uma métrica unificada a ser usada para ranquear a recomendação por relevância. Dentre as áreas de aplicação dos SR, destaca-se a recomendação de notícias, sendo utilizada por um público heterogêneo, amplo e exigente por relevância. Neste contexto, a presente pesquisa apresenta uma abordagem híbrida para recomendação de notícias construída por meio de uma arquitetura implementada para provar os conceitos de um sistema de recomendação. Esta arquitetura foi validada por meio da utilização de um corpus de notícias e pela realização de um experimento online. Por meio do experimento foi possível observar a capacidade da arquitetura em relação aos requisitos de um sistema de recomendação de notícias e também confirmar a hipótese no que se refere à privilegiar recomendações com base em similaridade, popularidade, diversidade, novidade e serendipidade. Foi observado também uma evolução nos indicadores de leitura, curtida, aceite e serendipidade conforme o sistema foi acumulando histórico de preferências e soluções. Por meio da análise da métrica unificada para ranqueamento foi possível confirmar sua eficácia ao verificar que as notícias melhores colocadas no ranqueamento foram as mais aceitas pelos usuários / Recommendation Systems (RS) are software capable of suggesting items to users based on the history of user interactions or by similarity metrics that can be compared by item, user, or both. There are different types of RS and those which most interest in this work are content-based, knowledge-based and collaborative filtering. Achieving adequate results to user\'s expectations is a hard goal due to the inherent subjectivity of human behavior, thus, the RS need efficient and effective solutions to: modeling the data that will support the recommendation; the information retrieval that describes the data; combining this information within similarity, popularity or suitability metrics; creation of descriptive models of the items under recommendation; and evolution of the systems intelligence to learn from the user\'s interaction. Decision-making by a RS is a complex task that can be implemented according to the view of fields such as artificial intelligence and data mining. In the artificial intelligence field there are studies concerning the method of case-based reasoning that works with the principle that if something worked in the past, it may work again in a new similar situation the one in the past. The case-based recommendation works with structured items, represented by a set of attributes and their respective values (within a ``case\'\' model), providing known and adapted solutions. Data mining area can build descriptive models to RS and also handle, manipulate and analyze textual data, constituting one option to create elements to compose a recommendation. One way to minimize the weaknesses of an approach is to adopt aspects based on a hybrid solution, which in this work considers: taking advantage of the different types of RS; using problem-solving techniques; and combining resources from different sources to compose a unified metric to be used to rank the recommendation by relevance. Among the RS application areas, news recommendation stands out, being used by a heterogeneous public, ample and demanding by relevance. In this context, the this work shows a hybrid approach to news recommendations built through a architecture implemented to prove the concepts of a recommendation system. This architecture has been validated by using a news corpus and by performing an online experiment. Through the experiment it was possible to observe the architecture capacity related to the requirements of a news recommendation system and architecture also related to privilege recommendations based on similarity, popularity, diversity, novelty and serendipity. It was also observed an evolution in the indicators of reading, likes, acceptance and serendipity as the system accumulated a history of preferences and solutions. Through the analysis of the unified metric for ranking, it was possible to confirm its efficacy when verifying that the best classified news in the ranking was the most accepted by the users

Sistemas de Recomendação (SR) são softwares capazes de sugerir itens aos usuários com base no histórico de interações de usuários ou por meio de métricas de similaridade que podem ser comparadas por item, usuário ou ambos. Existem diferentes tipos de SR e dentre os que despertam maior interesse deste trabalho estão: SR baseados em conteúdo; SR baseados em conhecimento; e SR baseado em filtro colaborativo. Alcançar resultados adequados às expectativas dos usuários não é uma meta simples devido à subjetividade inerente ao comportamento humano, para isso, SR precisam de soluções eficientes e eficazes para: modelagem dos dados que suportarão a recomendação; recuperação da informação que descrevem os dados; combinação dessas informações dentro de métricas de similaridade, popularidade ou adequabilidade; criação de modelos descritivos dos itens sob recomendação; e evolução da inteligência do sistema de forma que ele seja capaz de aprender a partir da interação com o usuário. A tomada de decisão por um sistema de recomendação é uma tarefa complexa que pode ser implementada a partir da visão de áreas como inteligência artificial e mineração de dados. Dentro da área de inteligência artificial há estudos referentes ao método de raciocínio baseado em casos e da recomendação baseada em casos. No que diz respeito à área de mineração de dados, os SR podem ser construídos a partir de modelos descritivos e realizar tratamento de dados textuais, constituindo formas de criar elementos para compor uma recomendação. Uma forma de minimizar os pontos fracos de uma abordagem, é a adoção de aspectos baseados em uma abordagem híbrida, que neste trabalho considera-se: tirar proveito dos diferentes tipos de SR; usar técnicas de resolução de problemas; e combinar recursos provenientes das diferentes fontes para compor uma métrica unificada a ser usada para ranquear a recomendação por relevância. Dentre as áreas de aplicação dos SR, destaca-se a recomendação de notícias, sendo utilizada por um público heterogêneo, amplo e exigente por relevância. Neste contexto, a presente pesquisa apresenta uma abordagem híbrida para recomendação de notícias construída por meio de uma arquitetura implementada para provar os conceitos de um sistema de recomendação. Esta arquitetura foi validada por meio da utilização de um corpus de notícias e pela realização de um experimento online. Por meio do experimento foi possível observar a capacidade da arquitetura em relação aos requisitos de um sistema de recomendação de notícias e também confirmar a hipótese no que se refere à privilegiar recomendações com base em similaridade, popularidade, diversidade, novidade e serendipidade. Foi observado também uma evolução nos indicadores de leitura, curtida, aceite e serendipidade conforme o sistema foi acumulando histórico de preferências e soluções. Por meio da análise da métrica unificada para ranqueamento foi possível confirmar sua eficácia ao verificar que as notícias melhores colocadas no ranqueamento foram as mais aceitas pelos usuários / Recommendation Systems (RS) are software capable of suggesting items to users based on the history of user interactions or by similarity metrics that can be compared by item, user, or both. There are different types of RS and those which most interest in this work are content-based, knowledge-based and collaborative filtering. Achieving adequate results to user\'s expectations is a hard goal due to the inherent subjectivity of human behavior, thus, the RS need efficient and effective solutions to: modeling the data that will support the recommendation; the information retrieval that describes the data; combining this information within similarity, popularity or suitability metrics; creation of descriptive models of the items under recommendation; and evolution of the systems intelligence to learn from the user\'s interaction. Decision-making by a RS is a complex task that can be implemented according to the view of fields such as artificial intelligence and data mining. In the artificial intelligence field there are studies concerning the method of case-based reasoning that works with the principle that if something worked in the past, it may work again in a new similar situation the one in the past. The case-based recommendation works with structured items, represented by a set of attributes and their respective values (within a ``case\'\' model), providing known and adapted solutions. Data mining area can build descriptive models to RS and also handle, manipulate and analyze textual data, constituting one option to create elements to compose a recommendation. One way to minimize the weaknesses of an approach is to adopt aspects based on a hybrid solution, which in this work considers: taking advantage of the different types of RS; using problem-solving techniques; and combining resources from different sources to compose a unified metric to be used to rank the recommendation by relevance. Among the RS application areas, news recommendation stands out, being used by a heterogeneous public, ample and demanding by relevance. In this context, the this work shows a hybrid approach to news recommendations built through a architecture implemented to prove the concepts of a recommendation system. This architecture has been validated by using a news corpus and by performing an online experiment. Through the experiment it was possible to observe the architecture capacity related to the requirements of a news recommendation system and architecture also related to privilege recommendations based on similarity, popularity, diversity, novelty and serendipity. It was also observed an evolution in the indicators of reading, likes, acceptance and serendipity as the system accumulated a history of preferences and solutions. Through the analysis of the unified metric for ranking, it was possible to confirm its efficacy when verifying that the best classified news in the ranking was the most accepted by the users

Information Filtering and Recommender Systems have been used and has been implemented in various ways from various entities since the dawn of the Internet, and state-of-the-art approaches rely on Machine Learning and Deep Learning in order to create accurate and personalized recommendations for users in a given context. These models require big amounts of data with a variety of features such as time, location and user data in order to find correlations and patterns that other classical models such as matrix factorization and collaborative filtering cannot. This thesis researches, implements and compares a variety of models with the primary focus of Machine Learning and Deep Learning for the task of music recommendation and do so successfully by representing the task of recommendation as a multi-class extreme classification task with 100 000 distinct labels. By comparing fourteen different experiments, all implemented models successfully learn features such as time, location, user features and previous listening history in order to create context-aware personalized music predictions, and solves the cold start problem by using user demographic information, where the best model being capable of capturing the intended label in its top 100 list of recommended items for more than 1/3 of the unseen data in an offine evaluation, when evaluating on randomly selected examples from the unseen following week. / Informationsfiltrering och rekommendationssystem har använts och implementeratspå flera olika sätt från olika enheter sedan gryningen avInternet, och moderna tillvägagångssätt beror påMaskininlärrning samtDjupinlärningför att kunna skapa precisa och personliga rekommendationerför användare i en given kontext. Dessa modeller kräver data i storamängder med en varians av kännetecken såsom tid, plats och användardataför att kunna hitta korrelationer samt mönster som klassiska modellersåsom matris faktorisering samt samverkande filtrering inte kan. Dettaexamensarbete forskar, implementerar och jämför en mängd av modellermed fokus påMaskininlärning samt Djupinlärning för musikrekommendationoch gör det med succé genom att representera rekommendationsproblemetsom ett extremt multi-klass klassifikationsproblem med 100000 unika klasser att välja utav. Genom att jämföra fjorton olika experiment,så lär alla modeller sig kännetäcken såsomtid, plats, användarkänneteckenoch lyssningshistorik för att kunna skapa kontextberoendepersonaliserade musikprediktioner, och löser kallstartsproblemet genomanvändning av användares demografiska kännetäcken, där den bästa modellenklarar av att fånga målklassen i sin rekommendationslista medlängd 100 för mer än 1/3 av det osedda datat under en offline evaluering,när slumpmässigt valda exempel från den osedda kommande veckanevalueras.

M. Tech. (Department of Information and Communication Technology, Faculty of Applied and Computer Sciences), Vaal University of Technology. / Since their emergence in the 1990’s, recommendation systems have transformed the intelligence of both the web and humans. A pool of research papers has been published in various domains of recommendation systems. These include content based, collaborative and hybrid filtering recommendation systems. Recommendation systems suggest items to users and their principal purpose is to increase sales and recommend items that are predicted to be suitable for users. They achieve this through making calculations based on data that is available on the system. In this study, we give evidence that the research on group recommendation systems must look more carefully at the dynamics of group decision-making in order to produce technologies that will be more beneficial for groups based on the individual interests of group members while also striving to maximise satisfaction. The matrix factorization algorithm of collaborative filtering was used to make predictions and three movie recommendation for each and every individual user. The three recommendations were of three highest predicted movies above the pre-set threshold which was three. Thereafter, four virtual groups of varied sizes were formed based on four highest predicted movies of the users in the dataset. Plurality voting strategy was used to achieve this. A publicly available dataset based on Group Recommender Systems Enhanced by Social Elements, constructed by Lara Quijano from the Group of Artificial Intelligence Applications (GIGA), was used for experiments. The developed recommendation system was able to successfully make individual movie recommendations, generate virtual groups, and recommend movies to these respective groups. The system was evaluated for accuracy in making predictions and it was able to achieve 0.7027 MAE and 0.8996 RMSE. This study was able to recommend to virtual groups to enable social network group members to engage in discussions of recommended items. The study encourages members in engaging in similar activities in their respective physical locations and then discuss on social network.

在人們的日常生活中，推薦是很普遍的一種社會行為，它使人們不必親自去體驗所有的事物，可透過別人的經驗來得知一件事情或商品的好或壞。隨著科技的快速發展與網際網路的普及，電子商務已逐漸的融入社會，成為人類生活中不可或缺的一部分。然而在網路上過量的資訊，使得個人在資訊的使用與搜尋上面臨極大的挑戰，更加刺激了對於推薦資訊的需求，因此許多推薦技術相繼提出，推薦系統也應運而生，不僅使得推薦的範圍擴大了，推薦的型態也更為豐富多元；同時，在近年電子商務的發展中，對於個人化與顧客導向服務的愈益重視，使得推薦系統逐漸成為一種必要的線上服務。 在眾多的推薦技術之中，協同過濾推薦方法是最成功且最常被採用的推薦技術之一，許多台灣的拍賣平台上也都有採用類似概念的推薦系統，像是Yahoo!拍賣、露天拍賣上的評價機制均屬此類。然而，現行的拍賣評價機制都沒有採用社會網路的技術，本研究希望透過協同過濾與社會網路的結合，讓評價機制更趨於完備。 本研究以台灣最大的拍賣網站Yahoo!為例，蒐集了44萬筆交易記錄，並以推薦網(ReferralWeb)系統的矩陣方法為基礎，找出人與商品的關係、商品與類別的關係、人與人的關係，建立起一個社會網路，讓使用者可查詢特定領域的專家，並與之交易。除此之外，也可直接詢問專家關於商品的資訊或購買技巧。透過這樣的機制，希望能降低消費者在購買商品時所產生的交易糾紛，讓人們在網路上的購物體驗能變得更好。 / Nowadays, recommendation is a common social behavior between people. People can evaluate things or commodities from others’ experience and opinions instead of their own experiences. Along with the development of technology and Internet today, E-commerce has become an indispensable part of human life. However, due to the overloaded information, people face a fantastic challenge when accessing and searching on the Internet. Therefore, many methods of recommendation were proposed, and systems of recommendation are to come with the tide of fashion. In addition, the development of E-commerce emphasized on personalization and customer-oriented services more in recent years, which make recommendation system becomes a necessary on-line service gradually. Collaborative Filtering is the most successful and adopted one in numerous recommendation methods. There are many auction platforms in Taiwan also use recommendation systems, such like "Yahoo Auction", "Ruten Auction", etc. However, the previous mentioned recommendation mechanisms haven’t used Social Network technology; this study will propose an recommendation system which combines Collaborative Filtering and Social Network technology. This research collects 440,000 transaction data from the Yahoo auction platform, which is the biggest auction website in Taiwan. Based on the matrix method of ReferralWeb system(Shah, 1997), this research would like to build up the matrix of relationships between Person-Commodity, Commodity-Category, and Person-Person. Based on the three matrixes, finally builds up a Social Network. In the Social Network, users can enquire experts refer to the specific category of commodity, and then refer to the shops which the experts like or directly ask them the commodity information and purchase skill. Relying on the mechanism proposed by this research, our goals are to reduce the transaction disputes arising from consumers purchase commodities, and to let people have better experiences in on-line shopping.

Avec l’émergence du Web de données, notamment les données ouvertes liées, une abondance de données est devenue disponible sur le web. Cependant, les ensembles de données LOD et leurs sous-graphes inhérents varient fortement par rapport a leur taille, le thème et le domaine, les schémas et leur dynamicité dans le temps au niveau des données. Dans ce contexte, l'identification des jeux de données appropriés, qui répondent a des critères spécifiques, est devenue une tâche majeure, mais difficile a soutenir, surtout pour répondre a des besoins spécifiques tels que la recherche d'entités centriques et la recherche des liens sémantique des données liées. Notamment, en ce qui concerne le problème de liage des données, le besoin d'une méthode efficace pour la recommandation des jeux de données est devenu un défi majeur, surtout avec l'état actuel de la topologie du LOD, dont la concentration des liens est très forte au niveau des graphes populaires multi-domaines tels que DBpedia et YAGO, alors qu'une grande liste d'autre jeux de données considérés comme candidats potentiels pour le liage est encore ignorée. Ce problème est dû a la tradition du web sémantique dans le traitement du problème de "identification des jeux de données candidats pour le liage". Bien que la compréhension de la nature du contenu d'un jeu de données spécifique est une condition cruciale pour les cas d'usage mentionnées, nous adoptons dans cette thèse la notion de "profil de jeu de données"- un ensemble de caractéristiques représentatives pour un jeu de données spécifique, notamment dans le cadre de la comparaison avec d'autres jeux de données. Notre première direction de recherche était de mettre en œuvre une approche de recommandation basée sur le filtrage collaboratif, qui exploite à la fois les prols thématiques des jeux de données, ainsi que les mesures de connectivité traditionnelles, afin d'obtenir un graphe englobant les jeux de données du LOD et leurs thèmes. Cette approche a besoin d'apprendre le comportement de la connectivité des jeux de données dans le LOD graphe. Cependant, les expérimentations ont montré que la topologie actuelle de ce nuage LOD est loin d'être complète pour être considéré comme des données d'apprentissage.Face aux limites de la topologie actuelle du graphe LOD, notre recherche a conduit a rompre avec cette représentation de profil thématique et notamment du concept "apprendre pour classer" pour adopter une nouvelle approche pour l'identification des jeux de données candidats basée sur le chevauchement des profils intensionnels entre les différents jeux de données. Par profil intensionnel, nous entendons la représentation formelle d'un ensemble d'étiquettes extraites du schéma du jeu de données, et qui peut être potentiellement enrichi par les descriptions textuelles correspondantes. Cette représentation fournit l'information contextuelle qui permet de calculer la similarité entre les différents profils d'une manière efficace. Nous identifions le chevauchement de différentes profils à l'aide d'une mesure de similarité semantico-fréquentielle qui se base sur un classement calcule par le tf*idf et la mesure cosinus. Les expériences, menées sur tous les jeux de données lies disponibles sur le LOD, montrent que notre méthode permet d'obtenir une précision moyenne de 53% pour un rappel de 100%.Afin d'assurer des profils intensionnels de haute qualité, nous introduisons Datavore- un outil oriente vers les concepteurs de métadonnées qui recommande des termes de vocabulaire a réutiliser dans le processus de modélisation des données. Datavore fournit également les métadonnées correspondant aux termes recommandés ainsi que des propositions des triples utilisant ces termes. L'outil repose sur l’écosystème des Vocabulaires Ouverts Lies (LOV) pour l'acquisition des vocabulaires existants et leurs métadonnées. / With the emergence of the Web of Data, most notably Linked Open Data (LOD), an abundance of data has become available on the web. However, LOD datasets and their inherent subgraphs vary heavily with respect to their size, topic and domain coverage, the schemas and their data dynamicity (respectively schemas and metadata) over the time. To this extent, identifying suitable datasets, which meet specific criteria, has become an increasingly important, yet challenging task to supportissues such as entity retrieval or semantic search and data linking. Particularlywith respect to the interlinking issue, the current topology of the LOD cloud underlines the need for practical and efficient means to recommend suitable datasets: currently, only well-known reference graphs such as DBpedia (the most obvious target), YAGO or Freebase show a high amount of in-links, while there exists a long tail of potentially suitable yet under-recognized datasets. This problem is due to the semantic web tradition in dealing with "finding candidate datasets to link to", where data publishers are used to identify target datasets for interlinking.While an understanding of the nature of the content of specific datasets is a crucial prerequisite for the mentioned issues, we adopt in this dissertation the notion of "dataset profile" - a set of features that describe a dataset and allow the comparison of different datasets with regard to their represented characteristics. Our first research direction was to implement a collaborative filtering-like dataset recommendation approach, which exploits both existing dataset topic proles, as well as traditional dataset connectivity measures, in order to link LOD datasets into a global dataset-topic-graph. This approach relies on the LOD graph in order to learn the connectivity behaviour between LOD datasets. However, experiments have shown that the current topology of the LOD cloud group is far from being complete to be considered as a ground truth and consequently as learning data.Facing the limits the current topology of LOD (as learning data), our research has led to break away from the topic proles representation of "learn to rank" approach and to adopt a new approach for candidate datasets identication where the recommendation is based on the intensional profiles overlap between differentdatasets. By intensional profile, we understand the formal representation of a set of schema concept labels that best describe a dataset and can be potentially enriched by retrieving the corresponding textual descriptions. This representation provides richer contextual and semantic information and allows to compute efficiently and inexpensively similarities between proles. We identify schema overlap by the help of a semantico-frequential concept similarity measure and a ranking criterion based on the tf*idf cosine similarity. The experiments, conducted over all available linked datasets on the LOD cloud, show that our method achieves an average precision of up to 53% for a recall of 100%. Furthermore, our method returns the mappings between the schema concepts across datasets, a particularly useful input for the data linking step.In order to ensure a high quality representative datasets schema profiles, we introduce Datavore| a tool oriented towards metadata designers that provides rankedlists of vocabulary terms to reuse in data modeling process, together with additional metadata and cross-terms relations. The tool relies on the Linked Open Vocabulary (LOV) ecosystem for acquiring vocabularies and metadata and is made available for the community.