A trend based investment decision approach using clustering and heuristic algorithm

Abstract

This paper proposes a stock price trend clustering and trend investment decision model by using a genetic algorithm to search for optimal solutions and the best investment strategies for different stock price trends. The new price trend clustering model identifies three types of stock price movements: uptrends, sideways trends, and downtrends. Unfortunately, trends discovered through stock price movements or technical indicator graphs are typically subjective and unquantifiable. This paper takes daily stock prices and trading volume data from the China Shanghai Stock Exchange Composite Index (SSECI) from January 2, 1997 to August 31, 2012, to examine the performance of the proposed trend clustering model. The proposed model is also compared to other popular stock market investment strategies to verify its validity. Research result shows that the proposed trend clustering model correctly identifies three different trends in the stock market. Furthermore, the trend investment strategy model developed by using genetic algorithm methodology performs better than other investment strategies, namely, Granville's rule, the KD indicator strategy, the buys and holds strategy, and GMA rules, in both bull and bear market periods. Research results prove the proposed new model to be a stable and valid investment strategy.