Selecting Stocks Using A Genetic Algorithm: A Case of Real Estate Investment Trusts (REITs)

Abstract

     Many investors invest according to analysts' recommendations, but some studies detect sizable abnormal returns associated with those recommendations. A tool that imitates the mechanism of the recommendation process should enable individual investors to formulate a profitable investment strategy and to select stocks that analysts have not covered. This paper presents a genetic algorithm (GA) method for identifying attractive stocks. The GA method has shown promising performance as a tool with which investors can develop a profitable investment strategy. GA is able to pick stocks with a "buy" rating at 62.50%, compared to 36.45% for stocks picked at random.