Sells Out or Piles Up? A Sentiment Autoregressive Model for Predicting Sales Performance

摘要

The development of e-commerce has witnessed the explosion of online reviews which represent the voices of the public. These reviews are helpful for consumers in making purchasing decisions, and this effect can be observed by some easy-to-measure economic variables, such as sales performance or product prices. In this paper, we study the problem of mining sentiment information from reviews and investigate whether applying sentiment analysis methods can turn out better sales predictions. Based on the nature of various presentations of sentiments, we propose a Latent Sentiment Language (LSL) Model to address this challenge, in which sentiment-language model and sentiment-LDA are used to capture the explicit and implicit sentiment information respectively. Subsequently, we explore ways to use such information to predict product sales, and to generate an SAR, a sentiment autoregressive model. Extensive experiments indicate the predictive power of sentiment information, as well as the superior performance of the SAR model.