Benefit Segmentation of Online Customer Reviews Using Random Forest
Kenjiro Torizuka, H. Oi, Fumiaki Saıtoh, Shinta Ishizu · 2018
The purpose of this study is to propose a new benefit segmentation method based on customer reviews existing on the web. With the diversification in customer needs, it is difficult to accurately identify the needs of customers with market segmentation using demographic information. Therefore, it is important in marketing to segment the customer market based on the benefits that customers receive for products or services. In this research, we use the random forest algorithm for benefit segmentation, as this algorithm identifies training data with high accuracy even if noise and outliers exist, and it is widely used for analysis of text data. In our experiment, we analyzed customer reviews for hotels. We treated the reason for using hotels as the benefit, and analyzed topics based on word frequency in the text data as explanatory variables. We extracted factors that influenced each benefit to determine customer needs.