LG-trader: Stock Trading Decision Support Based On Feature Selection By Weighted Localized Generalization Error Model

Stock trading is an important financial activity of human society. Machinelearning techniques are adopted to provide trading decision support bypredicting the stock price or trading signals of the next day. Decisions are usuallymade by analyzing technical indices and fundamental analysis of companies.There are two major machine learning research problems for stock tradingdecision support: classifier architecture selection and feature selection. In thiswork, we propose the LG-Trader which will deal with these two problemssimultaneously using a genetic algorithm minimizing a new Weighted LocalizedGeneralization Error (wL-GEM). An issue being ignored in current machinelearning based stock trading researches is the imbalance among buy, hold andsell decisions. Usually hold decision is the majority in comparison to both buyand sell decisions. So, the wL-GEM is proposed to balance classes by penalizingheavier for generalization error being made in minority classes. Moreover, thefeature selection based on wL-GEM helps to select most useful technical indicesamong choices for each stock. Experimental results demonstrate that theLG-Trader yields higher profits and rates of return in both stock and indextrading.