Opinion mining in social media as a tool to measure consumer (in)satisfaction

• Autor(a) principal: Sancliment Iglesias, Luis
• Data de Publicação: 2021
• Outros Autores: Fukumi Tsunoda, Denise
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista Brasileira de Educação em Ciência da Informação
• Texto Completo: https://portal.abecin.org.br/rebecin/article/view/235
• topic: Desicion tree; Sentiment analysis; Text mining; Data mining.; Árbol de decisión; Análisis de sentimientos; Minería de textos; Minería de datos.; Árvore de decisão; Análise de sentimento; Mineração de Texto; Mineração de dados.

A quantitative study that aims to explore the contribution of opinion mining in databases extracted from Facebook for the measurement of consumer (in)satisfaction. It aims to propose a flow that assists in the steps of the process of knowledge discovery in text and select tools for opinion mining at the sentence level, where the positive, negative and neutral sentiment is analyzed. A car brand database extracted from Facebook with four pre-processing treatments is submitted to the proposed flow. The Naïve Bayes, SMO and J48 algorithms in the Weka tool are used for the processing stage. It presents satisfactory results in opinion mining with the best hit rate obtained using the SMO algorithm.It proposes future work in CSC (Consumer Care Service) databases with the application of this developed methodology and studies to discover the causes of consumer (in)satisfaction found in CSC and CSC 2.0 databases.