Analyzing the discriminative attributes of products using text mining focused on cosmetic reviews |
| |
| Authors: | Sung Guen Kim Juyoung Kang |
| |
| Institution: | 1. Department of MIS, Business School of Ajou University, San 5, Woncheon-dong, Yeongtong-gu, Suwon 443-749, South Korea;2. Department of e-Business, Business School of Ajou University, San 5, Woncheon-dong, Yeongtong-gu, Suwon 443-749, South Korea |
| |
| Abstract: | Consumers evaluate products through online reviews, in addition to sharing their product experiences. Online reviews affect product marketing, and companies use online reviews to investigate consumer attitudes and perceptions of their products. However, when analyzing a review, it is often the case that specific contexts are not taken into consideration and meaningful information is not obtained from the analysis results. This study suggests a methodology for analyzing reviews in the context of comparing two competing products. In addition, by analyzing the discriminative attributes of competing products, we were able to derive more specific information than an overall product analysis. Analyzing the discriminative attributes in the context of comparing competing products provides clarity on analyzing the strengths and weaknesses of competitive products and provides realistic information that can help the company's management activities. Considering this purpose, this study collected a review of the BB Cream product line in the cosmetics field. The analysis was sequentially carried out in three stages. First, we extracted words that represent discriminative attributes by analyzing the percentage difference of words. Second, different attribute words were classified according to the meaning used in the review by using latent semantic analysis. Finally, the polarity of discriminative attribute words was analyzed using Labeled-LDA. This analysis method can be used as a market research method as it can extract more information than a traditional survey or interview method, and can save cost and time through the automation of the program. |
| |
| Keywords: | Text mining Latent semantic analysis LSA Labeled latent Dirichlet allocation Discriminative attributes of products |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
