New histogram-based user and item profiles for recommendation systems

• Autor(a) principal: SAMPAIO NETO, Delmiro Daladier
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Institucional da UFPE
• Texto Completo: https://repositorio.ufpe.br/handle/123456789/44841
• topic: Inteligência computacional; Sistemas de recomendação; Dados simbólicos; Histogramas

Recommendation systems play an important role in businesses such as e-commerce, digital entertainment and online education. Most recommendation systems are implemented using numerical or categorical data, that is, traditional data. This type of data can be a limiting factor when used to model complex concepts where there is internal variability or internal structure in the data. In order to overcome these limitations, symbolic data are used, where values can be intervals, probability distributions or lists of values. Symbolic data can benefit recommendation systems and this work introduces a methodology to construct recommendation systems using symbolic descriptions for users and items. The proposed methodology can be applied in the implementation of recommendation systems based on content or based on collaborative filtering. In the content-based approach, user profiles and item profiles are created from symbolic descriptions of their features and a list of items are matched against a user profile. In the approach based on collaborative filtering, user profiles are built and users are grouped to form a neighborhood, products rated by users of this neighborhood are recommended based on the similarity between the neighbor and the user who will receive the recommendation. Experiments are carried out to evaluate the effectiveness of the methodology proposed in this work in relation to existing methodologies in the literature for the two recommendation system approaches. In the experiments, it was shown that the methodology proposed in this work is able to produce ranked lists with higher quality than the methodologies in the literature, i.e., lists where items with greater relevance appear in the first positions. A movie domain dataset is used in these experiments and their results show the usefulness of the proposed methodology.