This research is to develop an opinion mining application which allows users to clarify what the reviews on the laptop mentioned. The aim of the research is to analyze user’s opinions from laptop reviews on popular online communities. The proposed methodology is composed of four essential processes: preparing data for analysis, detecting subjective text paragraphs, identifying the aspects and classifying the sentiments of text paragraphs. The subjective textual contents are determined by detecting subjective words occurred in the sentences of text paragraphs. Then, only the subjective paragraphs might be classified into specific aspects using comparisons with the vocabularies of aspect domains. Finally, the paragraph sentiments will be categorized into positive or negative opinions using the Naïve Bayes classifier. The experimental results with the performance evaluation showed that the accuracy and precision of the subjective detection of text paragraphs are greater than 90%. In addition, the accuracy and precision of sentiment classification are more than 70%. Therefore, this tool can help consumers in categorizing laptop review paragraphs into aspects and sentiment groups for making selections before purchasing laptops.
Keywords: opinion mining; review analysis; laptop reviews; Naïve Bayes
*Corresponding author: Tel..: +66 22 18 5170 Fax: +66 22 55 2287
E-mail: pakawan.p@chula.ac.th
Pugsee*, P. undefined. ., & Chatchaithanawat, T. undefined. . (2020). Opinion Mining for Laptop Reviews using Naïve Bayes. CURRENT APPLIED SCIENCE AND TECHNOLOGY, 278-294.
