STOCK PRICE MOVEMENTS PREDICTION WITH TEXTUAL INFORMATION

摘要

This paper addresses the problem of predicting the direction of stock movements for Chinese stock markets. In terms of feature extraction, six technical indicators are used as inputs for our model. Apart from that, we crawled lots of financial news from Netease Finance website and extract the headlines embedding in a vector as an additional feature. The study compares three prediction models, support vector machine, L1 regularized logistic regression and XG Boost. Evaluation is carried out on historical data from January 2014 to October 2017 for five stocks. The experimental results show that the performance of all the prediction models is improved when those models add textual information features.