Distância de Barbieri: uma métrica para identificar similaridade entre perfis de consumidores

This work is part of the study of recommender systems with content-based filtering, having as motivation the observation of user behavior in an Enterprise Resource Planning (ERP). The main contribution of the work is the development of Barbieri Distance, a metric whose purpose is to measure the simi...

Descripción completa

Detalles Bibliográficos
Autores: Barbieri, Luiz Eugênio, Cervi, Cristiano Roberto
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2020
País:Brasil
Institución:Sociedade Brasileira de Computação (SBC)
Repositorio:Brazilian Journal of Information Systems
Idioma:portugués
OAI Identifier:oai:journals-sol.sbc.org.br:article/537
Acceso en línea:https://journals-sol.sbc.org.br/index.php/isys/article/view/537
Access Level:acceso abierto
Palabra clave:Content-based filtering
Levenshtein distance
Recommender systems
Similarity
Distância Levenshtein
Filtragem baseada em conteúdo
Similaridade
Sistema de recomendação
Descripción
Sumario:This work is part of the study of recommender systems with content-based filtering, having as motivation the observation of user behavior in an Enterprise Resource Planning (ERP). The main contribution of the work is the development of Barbieri Distance, a metric whose purpose is to measure the similarity between buyers based on their purchase history. The metric is for situations where there is no buyer valuation data for the product purchased. Since it does not require ratings for items, because similarity happens when buyers buy too much or too little of the same product, it is possible to identify the similarity of the consumer profile based on their purchase history. In order to perform the metric validation experiments, a comparison method between buyer profiles is used, which presented satisfactory results in the calculation of similarity.