Design and Implementation of Machine Learning Based Multi Factor Quantitative Trading Strategy

Quantitative trading is a trading method that combines finance, mathematics, and computer science to achieve a goal. This method can help investors to filter out negative emotional influences effectively so that it is becoming more and more widely used in the Chinese stock market. Traditional quantitative trading strategies predict the trend of stock prices by analyzing fundamental indicators or technical indicators and building formulas quantitatively. However, this paper will use the emerging machine learning technologies to analyze the influence of multiple factors which impacts the stock price, then predict the return of particular stocks and stress the trading strategy. This research’s main work consists of obtaining financial data from third-party platforms and defining indicators; using Support Vector Machine, Random Forest, and XGBoost machine learning algorithms to build prediction models to predict which stock can bring excess return; generating the stock holding list; and designing the trading strategy accordingly. The outcomes are a multiple factors quantitative trading strategy based on machine learning which brings a steady excess return ratio while bearing low risk. The research achievement solves some problems in the current quantitative trading strategy: the selection of indicators is biased, a single machine learning model is not effective through a long period of time, the process of strategy research is not convenient enough.