Research on Short-Term Stock Price Trends Based on Machine Learning

Abstract

With the increasing emphasis on economic development, China’s economy is growing steadily and significantly, and the stock market is becoming more diversified and richer in products and derivatives, resulting in a larger and larger base of investors in the Chinese stock market over the years, with more significant changes in the investment environment of the primary stock market. In the analysis and research around stocks, the fitting and prediction of stock price changes have been one of the keys in the field of stock analysis, however, the current models and solutions for stock price fitting and prediction have not been well received, and are lacking in terms of realistic operability and applicability. In recent years, machine learning and its related models and methods have been widely used in the financial field, which has also promoted the development of stock price fitting forecasting. In order to further improve the accuracy of stock price fitting prediction, this paper introduces the multi-factor prediction model in traditional quantitative stock analysis into the stock price fitting prediction method and improves the general stock price fitting prediction method. This paper finds that the screening of factors that can significantly affect stock prices can indeed be accomplished by using the methodological properties of GBDT and FFM.