User demand analysis based on online review mining for product quality improvement

Online shopping has been integrated into the lifestyle of people in the contemporary society. Online comments made by users after shopping contain rich information. How to extract valuable information from these massive online review data has become a research hotspot of many experts and scholars. A novel Kano model of the classification indicator system based on online review mining is proposed in this paper, which uses two dimensions of review content depth and user attention to divide the review data into four categories. Meanwhile, through the sentiment analysis, the evolutionary sentiment curve of the comments over time can be obtained. Lenovo R7000P laptop is the research object in this paper. According to novel Kano model based on online review mining, the number of must-be demand samples is 469, the number of one-dimensional demand samples is 308, the number of attractive demand samples is 1938, and the number of indifferent demand samples is 308. Finally, the LDA model is used to analyze the four types of data. And the topic word distribution is output.