Automatic Stock Trading Algorithms using Matlab

Abstract

Stock Trading is one of the most efficient and quickest way to make money. Investors can gain a huge profit with only a small investment by making correct decisions to enter or leave the stock market. However, decision-making is always a difficult. Since the decisions are made by human, no matter the experienced or inexperienced investors, their judgement can be adversely affected by their physical and psychological conditions, which is prone to bad decisions. In this project, an Automatic Stock Trading Algorithm combined with the Oscillation Box Theory and Support Vector Regression is developed. The algorithm is divided into two parts. The first part is a Support Vector Regression estimators which are used to forecast the highest/lowest closing price of a stock in b-day future. The second part is a Stock Trading Strategy which is based on the Oscillation Box Theory. The two estimators from the first part of the system will be used as the upper and lower bound of the oscillation box respectively. Trading Simulations on different stocks in Hong Kong Stock Market are performed to show the results.