Prediction of Development Prospect of Electric Vehicles in China by Using Natural Language Processing

Abstract

It is believed that people's comments on a certain product may affect its sales condition. In this paper, we propose a method to predict the sales of electric vehicles by analyzing people's comments on social media. We scrap user comments from a Chinese social media “Weibo” and try to predict the electric vehicle sales in China by using Natural Language Processing (NLP). Sentiment score, number of comments and likes, and keyword existence are treated as input indicators. We test linear regression, random forest, and gradient boosting algorithm during the experiment. The result shows that the model which using gradient boosting algorithm to predict the market share of electric vehicles has the best performance.