Online shopping has become an important part of people’s daily life with the rapid development of e-commerce. In some domains such as books, electronics, and CD/DVDs, online shopping has surpassed or even replaced the traditional shopping method. Compared with traditional retailing, e-commerce is information intensive. One of the key factors to succeed in e-business is how to facilitate the consumers’ approaches to discover a product. Conventionally a product search engine based on a keyword search or category browser is provided to help users find the product information they need. The general goal of a product search system is to enable users to quickly locate information of interest and to minimize users’ efforts in search and navigation. In this process human factors play a significant role. Finding product information could be a tricky task and may require an intelligent use of search engines, and a non-trivial navigation of multilayer categories. Searching for useful product information can be frustrating for many users, especially those inexperienced users. This dissertation focuses on developing a new visual product search system that effectively extracts the properties of unstructured products, and presents the possible items of attraction to users so that the users can quickly locate the ones they would be most likely interested in. We designed and developed a feature extraction algorithm that retains product color and local pattern features, and the experimental evaluation on the benchmark dataset demonstrated that it is robust against common geometric and photometric visual distortions. Besides, instead of ignoring product text information, we investigated and developed a ranking model learned via a unified probabilistic hypergraph that is capable of capturing correlations among product visual content and textual content. Moreover, we proposed and designed a fuzzy hierarchical co-clustering algorithm for the collaborative filtering product recommendation. Via this method, users can be automatically grouped into different interest communities based on their behaviors. Then, a customized recommendation can be performed according to these implicitly detected relations. In summary, the developed search system performs much better in a visual unstructured product search when compared with state-of-art approaches. With the comprehensive ranking scheme and the collaborative filtering recommendation module, the user’s overhead in locating the information of value is reduced, and the user’s experience of seeking for useful product information is optimized.

Every single day, lots of users actively participate in social media sites (e.g., Facebook, YouTube, Last.fm, Flicker, etc.) upload photos, videos, share bookmarks, write blogs and annotate/comment on content provided by others. With the recent proliferation of social media sites, users are overwhelmed by the huge amount of available content. Therefore, organizing and retrieving appropriate multimedia content is becoming an increasingly important and challenging task. This challenging task led a number of research communities to concentrate on social tagging systems (also known as folksonomy) that allow users to freely annotate their media items (e.g., music, images, or video) with any sort of arbitrary words, referred to as tags. Tags assist users to organize their own content, as well as to find relevant content shared by other users. In this thesis, we first analyze how useful a folksonomy is for improving personalized services such as tag recommendation, tag-based search and item annotation. We then propose two new algorithms for social media retrieval and tag recommendation respectively. The first algorithm computes the latent preferences of tags for users from other similar tags, as well as latent annotations of tags for items from other similar items. We then seamlessly map the tags onto items, depending on an individual user’s query, to find the most desirable content relevant to the user’s needs. The second algorithm improves tag-recommendation and item annotation by adapting the Katz measure, a path-ensemble based proximity measure, for the use in social tagging systems. In this algorithm we model folksonomy as a weighted, undirected tripartite graph. We then apply the Katz measure to this graph, and exploit it to provide personalized tag recommendation for individual users. We evaluate our algorithms on two real-world folksonomies collected from Last.fm and CiteULike. The experimental results demonstrate that the proposed algorithms improve the search and the recommendation performance, and obtain significant gains in cold start situations where relatively little information is known about a user or an item

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-01-12T14:22:55Z No. of bitstreams: 1 crystiamkellepereiraesilva.pdf: 5368190 bytes, checksum: 22e15248de5dbc282e6d4324b03dca64 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-01-25T17:15:27Z (GMT) No. of bitstreams: 1 crystiamkellepereiraesilva.pdf: 5368190 bytes, checksum: 22e15248de5dbc282e6d4324b03dca64 (MD5) / Made available in DSpace on 2016-01-25T17:15:27Z (GMT). No. of bitstreams: 1 crystiamkellepereiraesilva.pdf: 5368190 bytes, checksum: 22e15248de5dbc282e6d4324b03dca64 (MD5) Previous issue date: 2015-03-27 / Existem inúmeros recursos educacionais distribuídos em diferentes repositórios que abordam um conjunto amplo de assuntos e que possuem objetivos educacionais distintos. A escolha adequada desses recursos educacionais é um desafio para os usuários que desejam usá-los para a sua formação intelectual. Nesse contexto surgem os Sistemas de Recomendação para auxiliar os usuários nessa tarefa. Para que seja possível gerar recomendações personalizadas, torna-se importante identificar informações que ajudem a definir o perfil do usuário e auxiliem na identificação de suas necessidades e interesses. O uso constante e cada vez mais intenso de algumas ferramentas tecnológicas faz com que inúmeras informações a respeito do perfil, dos interesses, das preferências, da forma de interação e do comportamento do usuário possam ser identificadas em decorrência da interação espontânea que ocorre nesses sistemas. Esse é o caso, por exemplo, das redes socais. Neste trabalho é apresentada a proposta e o desenvolvimento de uma arquitetura capaz de extrair características do perfil e do contexto educacional dos usuários, através da rede social Facebook e realizar recomendações de recursos educacionais de forma individualizada e personalizada que sejam condizentes com essas características. A solução proposta é apoiada por técnicas de extração de informações e ontologias para a extração, definição e enriquecimento das características e interesses dos usuários. As técnicas de Extração de Informação foram aplicadas aos textos associados às páginas curtidas e compartilhadas por usuários nas suas redes sociais para extrair informação estruturada que possa ser usada no processo de recomendação de recursos educacionais. Já as ontologias foram usadas para buscar interesses relacionados aos temas extraídos. A recomendação é baseada em repositório de objetos de aprendizagem e em repositórios de dados ligados e é realizada dentro das redes sociais, aproveitando o tempo despendido pelos usuários nas mesmas. A avaliação da proposta foi feita a partir do desenvolvimento de um protótipo, três provas de conceito e um estudo de caso. Os resultados da avaliação mostraram a viabilidade e uma aceitação relevante por parte dos usuários no sentido de extrair informações sobre os seus interesses educacionais, geradas automaticamente da rede social Facebook, enriquecê-las, encontrar interesses implícitos e usar essas informações para recomendar recursos educacionais. Foi verificada também a possibilidade da recomendação de pessoas, permitindo a formação de uma rede de interesses em torno de um determinado tema, indicando aos usuários bons parceiros para estudo e pesquisa. / There are several educational resources distributed in different repositories that address to a wide range of subjects and have different educational goals. The proper choice of these educational resources is a challenge for users who want to use them for their intellectual development. In this context, recommendation systems may help users in this task.In order to be able to generate personalized recommendations, it is important to identify information that will help to define user profile and assist in identifying his/her needs and interests. The constant and ever-increasing use of some technological tools allows the identification of different information about profile, interests, preferences, interaction style and user behavior from the spontaneous interaction that occurs in these systems, as, for example, the social networks. This paper presents the proposal and the development of one architecture able to extract users´ profile characteristics and educational context, from the Facebook social network and recommend educational resources in individualized and personalized manner, consistent with these characteristics. The proposed solution is supported by Information Extraction Techniques and ontologies for the extraction, enrichment and definition of user characteristics and interests. The Information Extraction techniques were applied to texts associated with “LIKE” and shared user´s pages on his social networks to extract structured information that can be used in the recommendation process of educational resources, the ontologies were used to search to interests related to extracted subjects. The recommendation process is based on learning objects repositories and linked data repositories and is carried out within social networks, taking advantage of user time spent at the web. The proposal evaluation was made from the development of a prototype, three proofs of concept and a case study. The evaluation results show the viability and relevant users´ acceptance in order to extract information about their educational interests, automatically generated from the Facebook social network, enrich these information, find implicit interests and use this information to recommend educational resources. It was also validated the possibility of people recommendation, enabling the establishment of interest network, based on a specific subject, showing good partners to study and research.

Com a grande variedade de produtos e serviços disponíveis na Web, os usuários possuem, em geral, muita liberdade de escolha, o que poderia ser considerado uma vantagem se não fosse pela dificuldade encontrada em escolher o produto ou serviço que mais atenda a suas necessidades dentro do vasto conjunto de opções disponíveis. Sistemas de recomendação são sistemas que têm como objetivo auxiliar esses usuários a identificarem itens de interesse em um conjunto de opções. A maioria das abordagens de sistemas de recomendação foca em recomendar itens mais relevantes para usuários individuais, não levando em consideração o contexto dos usuários. Porém, em muitas aplicações é importante também considerar informações contextuais para fazer as recomendações. Por exemplo, um usuário pode desejar assistir um filme com a sua namorada no sábado à noite ou com os seus amigos durante um dia de semana, e uma locadora de filmes na Web pode recomendar diferentes tipos de filmes para este usuário dependendo do contexto no qual este se encontra. Um grande desafio para o uso de sistemas de recomendação sensíveis ao contexto é a falta de métodos para aquisição automática de informação contextual para estes sistemas. Diante desse cenário, neste trabalho é proposto um método para extrair informações contextuais do conteúdo de páginas Web que consiste em construir hierarquias de tópicos do conteúdo textual das páginas considerando, além da bag-of-words tradicional (informação técnica), também informações mais valiosas dos textos como entidades nomeadas e termos do domínio (informação privilegiada). Os tópicos extraídos das hierarquias das páginas Web são utilizados como informações de contexto em sistemas de recomendação sensíveis ao contexto. Neste trabalho foram realizados experimentos para avaliação do contexto extraído pelo método proposto em que foram considerados dois baselines: um sistema de recomendação que não considera informação de contexto e um método da literatura de extração de contexto implementado e adaptado para este mestrado. Além disso, foram utilizadas duas bases de dados. Os resultados obtidos foram, de forma geral, muito bons apresentando ganhos significativos sobre o baseline sem contexto. Com relação ao baseline que extrai informação contextual, o método proposto se mostrou equivalente ou melhor que o mesmo. / With the wide variety of products and services available on the web, it is difficult for users to choose the option that most meets their needs. In order to reduce or even eliminate this difficulty, recommender systems have emerged. A recommender system is used in various fields to recommend items of interest to users. Most recommender approaches focus only on users and items to make the recommendations. However, in many applications it is also important to incorporate contextual information into the recommendation process. For example, a user may want to watch a movie with his girlfriend on Saturday night or with his friends during a weekday, and a video store on the Web can recommend different types of movies for this user depending on his context. Although the use of contextual information by recommendation systems has received great focus in recent years, there is a lack of automatic methods to obtain such information for context-aware recommender systems. For this reason, the acquisition of contextual information is a research area that needs to be better explored. In this scenario, this work proposes a method to extract contextual information of Web page content. This method builds topic hierarchies of the pages textual content considering, besides the traditional bag-of-words, valuable information of texts as named entities and domain terms (privileged information). The topics extracted from the hierarchies are used as contextual information in context-aware recommender systems. By using two databases, experiments were conducted to evaluate the contextual information extracted by the proposed method. Two baselines were considered: a recommendation system that does not use contextual information (IBCF) and a method proposed in literature to extract contextual information (\\methodological\" baseline), adapted for this research. The results are, in general, very good and show significant gains over the baseline without context. Regarding the \"methodological\" baseline, the proposed method is equivalent to or better than this baseline.

A extração de metadados semânticos de vídeos digitais para uso em serviços de personalização é importante, já que o conteúdo é adaptado segundo as preferências de cada usuário. Entretanto, apesar de serem encontradas várias propostas na literatura, as técnicas de indexação automática são capazes de gerar informações semânticas apenas quando o domínio do conteúdo é restrito. Alternativamente, existem técnicas para a criação manual dessas informações por profissionais, contudo, são dispendiosas e suscetíveis a erros. Uma possível solução seria explorar anotações colaborativas dos usuários, mas tal estratégia provoca a perda de individualidade dos dados, impedindo a extração de preferências do indivíduo a partir da interação. Este trabalho tem como objetivo propor uma arquitetura de personalização que permite a indexação multimídia de modo irrestrito e barato, utilizando anotações colaborativas, mas mantendo-se a individualidade dos dados para complementar o perfil de interesses do usuário com conceitos relevantes. A multimodalidade de metadados e de preferências também é explorada na presente tese, fornecendo maior robustez na extração dessas informações, e obtendo-se uma maior carga semântica que traz benefícios às aplicações. Como prova de conceito, este trabalho apresenta dois serviços de personalização que exploram a arquitetura proposta, avaliando os resultados por meio de comparações com abordagens previamente propostas na literatura / The extraction of semantic information from digital video is important to be used on personalization services because the content is adapted according to each users preferences. However, although it is possible to find several approaches in the literature, automatic indexing techniques are able to generate semantic metadata only when the contents domain is restricted. Alternatively, this information can be created manually by professionals, but this activity is time-consuming and error-prone. A possible solution would be to explore collaborative users annotations, but such approach has the disadvantage of lacking the individuality of annotations, hampering the extraction of users preferences from the interaction. This work has the objective of proposing a generic personalization architecture that allows multimedia indexing procedures to be accomplished in a cheap and unrestricted way. Such architecture uses collaborative annotations, but keeps the individuality of the data in order to augment the users profile with relevant concepts. The multimodality of metadata and users preferences is also explored in this work, which provides robustness during the extraction of semantic information, bringing benefits to applications. This work also presents two personalization services that explore the proposed architecture, along with evaluations that compare the obtained results with previously proposed approaches

O manejo nutricional adequado é fator determinante na produtividade e na qualidade dos frutos de macieira. Dentre os métodos para diagnóstico nutricional das plantas, destacam-se o critério de faixa de suficiência e o sistema integrado de diagnose e recomendação (DRIS). Este trabalho teve por objetivo avaliar o DRIS como método de interpretação de resultados de análises de folhas de plantas de macieira, estabelecendo normas adequadas para a cultura, e compará-lo com o método de diagnose nutricional, baseado no critério de faixa de suficiência, atualmente utilizado no Sul do Brasil. Buscou-se determinar, também a melhor época de amostragem de folhas de macieira para a aplicação do método DRIS. O estudo foi realizado na região produtora de maçã dos Campos de Cima da Serra, no Rio Grande do Sul, e nas regiões do Alto Vale do Rio do Peixe e Planalto Serrano, em Santa Catarina, em 70 pomares selecionados quanto à produtividade e técnicas de manejo do pomar, onde foram obtidas informações sobre a produtividade, espaçamento, porta-enxertos e realizada a amostragem de folhas e solo. Foram determinadas as concentrações de nitrogênio, fósforo, potássio, cálcio, magnésio, boro, cobre, ferro, manganês e zinco nas amostras de folhas e os valores de pH e os teores de matéria orgânica, fósforo, potássio, cálcio e magnésio nas amostras de solos. Também foram utilizados resultados de experimentos de adubação potássica e nitrogenada para avaliar a eficiência dos métodos DRIS, bem como de resultados de sazonalidade de nutrientes em três cultivares de macieira, para avaliar a época adequada de coleta de folhas para o método DRIS. Os índices DRIS foram calculados utilizando-se dois critérios para a escolha da ordem da razão dos nutrientes (Letzsch, 1985 e Walworth et al., 1986; Nick, 1998) e três formas de cálculo das funções dos nutrientes (Beaufils, 1973; Jones, 1981; Elwali & Gascho, 1984). Os resultados indicaram que: (i) A concentração dos nutrientes apresentou correlação positiva e significativa (p<0,01) com os respectivos índices DRIS, com exceção do N; (ii) O critério do “valor F” (Letzsch, 1985 e Walworth et al., 1986) mostrou-se mais eficiente que o “valor R” (Nick, 1998) para a escolha da ordem da razão dos nutrientes para a cultura da macieira; (iii) O Índice de Balanço Nutricional (IBN), calculado a partir das normas geradas, apresentou correlação negativa e significativa (p<0,01) com a produtividade para a população de referência, em todas as combinações de métodos testados; (iv) O método DRIS descrito por Elwali & Gascho (1984), utilizando o “valor F”, quando comparado com o critério de faixas de suficiência, apresentou diagnóstico nutricional mais eficiente que os demais métodos de cálculo do DRIS; (v) O método de cálculo do DRIS, com base no somatório das funções, descrito por Elwali & Gascho (1984), utilizando o “valor F” é o mais indicado para a cultura da macieira, por apresentar valores de IBN que melhor indicam o estado nutricional das plantas e pela eficiência no diagnóstico nutricional da cultura; (vi) A melhor época de coleta de folhas para o método DRIS esta situada entre a quinta e a décima quinta semana após a plena floração; (vii) As normas DRIS geradas neste trabalho foram adequadas para o diagnóstico nutricional da macieira, para as condições do Sul do Brasil. / The appropriate nutritional management is a decisive factor in fruit productivity and quality of apple trees. Among the several methods for nutritional diagnosis of the plants, the most important are the sufficiency range approach and the diagnosis and recommendation integrated system (DRIS). The objective of this work was evaluate DRIS as an interpretation method of results of analyses of apple tree leaves, establishing appropriate norms for the culture and comparing it with the sufficiency range approach currently used in the Southern Brazil, and determine the best sampling time of apple tree leaves for the application of the DRIS method. The study was carried out in the apple producing area of Campos de Cima da Serra (RS, Brazil), and in the areas of Alto Vale do Rio do Peixe and Planalto Serrano (SC, Brazil), in 70 orchards selected on basis of productivity and management techniques, where information on productivity, spacing, rootstock was obtained and leaf and soil sampling were performed. The concentrations of nitrogen, phosphorus, potassium, calcium, magnesium, boron, copper, iron, manganese and zinc were determined in the leaf samples as well as the pH values and the concentrations of organic matter, phosphorus, potassium, calcium and magnesium in the samples of soils. Results of fertilization experiments with potassium and nitrogen were also used to evaluate the efficiency of DRIS methods, as well as results of nutrient seasonally in three apple tree cultivars, to evaluate the appropriate leaf collection time for the DRIS method. The DRIS indices were calculated using two criteria to choose the order of the nutrient ratio (Letzsch, 1985 and Walworth et al., 1986; Nick, 1998) and three forms of calculating of the nutrient functions (Beaufils, 1973; Jones, 1981; Elwali & Gascho, 1984). The results indicated that: (i) the nutrient concentration presented positive and significant correlation (p<0.01) with the respective DRIS indices, except for N; (ii) the criterion of the "F value" (Letzsch, 1985 and Walworth et al., 1986) was shown to be more efficient than the "R value” (Nick, 1998) to choose the order of the nutrient ratio for apple tree culture; (iii) the Nutritional Balance Index (NBI), calculated from the generated norms, presented a negative and significant correlation (p <0.01) with productivity for the reference population in all combinations of methods tested; (iv) the DRIS method described by Elwali & Gascho (1984), using the "F value", when compared with the sufficiency range approach, presented a more efficient nutritional diagnosis than the other methods of DRIS calculation; (v) the method of DRIS calculation, based on the sum of the functions, described by Elwali & Gascho (1984), using the "F value" is the most suitable for apple tree culture, for presenting NBI values that best indicate the nutritional state of the plants and for the efficiency in the nutritional diagnosis of the culture; (vi) the best leaf sampling time for the DRIS method is between the fifth and the fifteenth week after full blossom; (vii) The DRIS norms generated in this work were appropriate for the nutritional diagnosis of apple trees, for the conditions of Southern Brazil.

The news landscape has changed during recent years because of the digitization. News can nowadays be found in both newspapers and on different sites online. The availability of the digital newspapers leads to competition among the news companies. To make the users stay on one specific platform for news, relevance is required in the content and oneway of creating relevance is through personalization, to tailor the content to each user. The focus of this thesis is therefore personalizing newspush notifications for a digital  newspaper and making them more relevant for users. The project was made in cooperation with VK Media, and their digital newspaper. The task in this thesis is to implement personalization of push notifications by building a recommendation system and to test the implemented system with data from VK. In order to perform the task, a dataset representing reading habits of VK’s users was extracted from their data warehouse. Then a user-based and content-based recommendation system was implemented in Python.The idea with the system is to recommend new articles that are sufficiently similar to one or more of the already read articles. Articles that may be liked by one of the most similar users should also be recommended. Finally, the system’s performance was evaluated with the data representing reading habits for VK’s users. The results show that the implemented system has better performance than the current solution without any personalization, when recommending a few articles to each user. The results from the evaluation also show that the more articles the users have read, the better predictions are possible to make. Thus, this thesis offers a first step towards meeting the expectations of more relevant content among VK’s users.

El texto completo de este trabajo no está disponible en el Repositorio Académico UPC por restricciones de la casa editorial donde ha sido publicado. / This article is about the problem and development of a fashion virtual assistant proposed by using a Kinect v2 camera and image processing, for fashion retail stores. It comes up mainly as a response to the inability of providing unique experiences during the shopping process through the use of diverse devices. Because of this, similar virtual assistant solutions, oriented to provide clothing recommendations, were analyzed to be able to provide software that could give a more personalized suggestion for the users based on their physical characteristics. / Revisión por pares

Many employment services base their online job recommendations to users based solely on explicit data in their profiles. The implicit data of what users for example click on, save and mark as irrelevant goes unused. Instead of making recommendations based on user behavior they make a direct comparison between user preferences and job ad attributes. A reason for this is the concern that the inclusion of implicit data can give odd recommendations resulting in a loss of credibility for the service. However, as research has shown this to be of great advantage to recommender systems. In this paper I implement a job recommender and test it both with user data including interaction history with job ads as well as with only explicit data. The results of the recommender with implicit data got better overall performance, but negligible gain in the ratio between true and false positives, or in other words the ratio between correct and incorrect recommendations.

[pt] A popularidade de plataformas para o armazenamento e compartilhamento de vídeo tem criado um volume massivo de horas de vídeo. Dado um conjunto de atores presentes em um vídeo, a geração de metadados com a determinação temporal dos intervalos em que cada um desses atores está presente, bem como a localização no espaço 2D dos quadros em cada um desses intervalos pode facilitar a recuperação de vídeo e a recomendação. Neste trabalho, nós investigamos a Clusterização Facial em Vídeo para a localização espaço-temporal de atores. Primeiro descrevemos nosso método de Clusterização Facial em Vídeo em que utilizamos métodos de detecção facial, geração de embeddings e clusterização para agrupar faces dos atores em diferentes quadros e fornecer a localização espaço-temporal destes atores. Então, nós exploramos, propomos, e investigamos aplicações inovadoras dessa localização espaço-temporal em três diferentes tarefas: (i) Reconhecimento Facial em Vídeo, (ii) Recomendação de Vídeos Educacionais e (iii) Posicionamento de Legendas em Vídeos 360 graus. Para a tarefa (i), propomos um método baseado na similaridade de clústeres que é facilmente escalável e obteve um recall de 99.435 por cento e uma precisão de 99.131 por cento em um conjunto de vídeos. Para a tarefa (ii), propomos um método não supervisionado baseado na presença de professores em diferentes vídeos. Tal método não requer nenhuma informação adicional sobre os vídeo e obteve um valor mAP aproximadamente 99 por cento. Para a tarefa (iii), propomos o posicionamento dinâmico de legendas baseado na localização de atores em vídeo 360 graus. / [en] The popularity of platforms for the storage and transmission of video content has created a substantial volume of video data. Given a set of actors present in a video, generating metadata with the temporal determination of the interval in which each actor is present, and their spatial 2D localization in each frame in these intervals can facilitate video retrieval and recommendation. In this work, we investigate Video Face Clustering for this spatio-temporal localization of actors in videos. We first describe our method for Video Face Clustering in which we take advantage of face detection, embeddings, and clustering methods to group similar faces of actors in different frames and provide the spatio-temporal localization of them. Then, we explore, propose, and investigate innovative applications of this spatio-temporal localization in three different tasks: (i) Video Face Recognition, (ii) Educational Video Recommendation and (iii) Subtitles Positioning in 360-video. For (i), we propose a cluster-matching-based method that is easily scalable and achieved a recall of 99.435 percent and precision of 99.131 percent in a small video set. For (ii), we propose an unsupervised method based on them presence of lecturers in different videos that does not require any additional information from the videos and achieved a mAP approximately 99 percent. For (iii), we propose a dynamic placement of subtitles based on the automatic localization of actors in 360-video.