• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Aprendizado neural de representa??o de conte?do para sistema de recomenda??o de filmes

Rassweiler Filho, Ralph Jos? 22 August 2017 (has links)
Submitted by Caroline Xavier (caroline.xavier@pucrs.br) on 2017-11-21T10:47:45Z No. of bitstreams: 1 DIS_RALPH_JOSE_RASSWEILER_FILHO_COMPLETO.pdf: 8289974 bytes, checksum: 5b62b872ae037f0894ab766c0251a4ef (MD5) / Approved for entry into archive by Caroline Xavier (caroline.xavier@pucrs.br) on 2017-11-21T10:48:02Z (GMT) No. of bitstreams: 1 DIS_RALPH_JOSE_RASSWEILER_FILHO_COMPLETO.pdf: 8289974 bytes, checksum: 5b62b872ae037f0894ab766c0251a4ef (MD5) / Made available in DSpace on 2017-11-21T10:48:13Z (GMT). No. of bitstreams: 1 DIS_RALPH_JOSE_RASSWEILER_FILHO_COMPLETO.pdf: 8289974 bytes, checksum: 5b62b872ae037f0894ab766c0251a4ef (MD5) Previous issue date: 2017-08-22 / Recommender systems are software used to generate personalized lists according to users profiles. The area is new and is growing since the internet popularization having its roots in information retrieval. Collaborative filtering is the most common approach of recommender systems used in both academy and industry because content-based filtering has problems such as lack of semantic information and poor content extraction techniques from items. Nowadays there are more content available in the form of multimedia such as video, images and text. Also, there are advances in pattern recognition though techniques like convolutional neural networks. In this work a convolutional neural network is used to extract features from movie trailers frames to further use these features to create a content-based recommender system with the goal of assessing whether the success of such networks on tasks like image classification and object detection also occur in the recommendation context. To evaluate that, the proposed method was compared with a media aesthetic detection method, two methods of feature extraction from text using TF-IDF and the traditional user and item collaborative filtering methods. Our results indicate that the proposed method is superior to the other content-based methods and is competitive to the collaborative filtering methods, being superior to the item-collaborative method regarding classification accuracy, and being superior to all other methods regarding execution time. In conclusion, we can state that the method using convolutional neural networks to represent items is promising for the recommender systems context. / Sistemas de recomenda??o s?o softwares cujo prop?sito ? gerar listas personalizadas, de acordo com as prefer?ncias de usu?rios. A ?rea ? bastante recente e est? em expans?o desde a populariza??o da internet tendo suas ra?zes em recupera??o de informa??o. Dos dois tipos tradicionais de sistemas de recomenda??o, a filtragem colaborativa ? a mais utilizada na academia e na ind?stria por trazer melhores resultados que o segundo tipo, a filtragem baseada em conte?do. Este ?ltimo sofre de problemas tais como a falta de informa??o sem?ntica e a dificuldade em extrair conte?do dos itens. Atualmente h? uma maior disponibilidade de conte?do de itens na forma de recursos multim?dia tais como v?deos, imagens e texto. Tamb?m houve avan?os no reconhecimento de padr?es em imagens atrav?s de t?cnicas como as redes neurais convolucionais. Neste trabalho, prop?e-se utilizar uma rede neural convolucional como extratora de atributos dos quadros que comp?e trailers de filmes que servem como base para um sistema de recomenda??o baseado em conte?do com o objetivo de avaliar se o sucesso destas redes em tarefas como classifica??o de imagens e detec??o de objetos tamb?m ocorre no contexto de recomenda??es. Para esta avalia??o, comparou-se o m?todo proposto com um m?todo de detec??o de est?tica de m?dia, dois m?todos de extra??o de conte?do de texto usando TF-IDF e os tradicionais m?todos colaborativos entre usu?rios e itens. Os resultados obtidos mostram que o m?todo proposto neste trabalho ? superior aos demais m?todos baseados em conte?do e ? competitivo com os m?todos colaborativos, superando o m?todo colaborativo entre itens na m?trica que representa acur?cia de classifica??o e tamb?m, superando todos os outros m?todos com rela??o ao tempo de execu??o. Concluiu-se que o m?todo que utiliza redes neurais convolucionais para representar itens ? promissor para o contexto de sistemas de recomenda??o.

Page generated in 0.1004 seconds