Predict Stock Price Trend by Using Classification Model

Abstract

The stock market is changing daily and people are paying more and more attention to stocks. After the establishment of the stock market, the research on stocks has become more and more influential. The core of researching stock is the stock future price trend, bullish or bearish. In order to predict stock information in simple and efficient ways, this paper aims to predict stock rise or fall by using classification model with better performance index and strong operability. Firstly, the paper acquires Maotai Corporations daily stock data from tushare package. To define the label up and down, the paper compares the daily closing price with its yesterday price. If it is positive, it is recorded as up; if it is negative, it is recorded as down. The random forest, logistic regression and SVM models are established respectively. The best model was selected by comparing three models evaluation scores. The results show that logistic regression is better than the other two models in predicting the rise and fall of stocks. This study can promote the cross integration of financial field and technical level and provide new ideas for future stock investment.