A DEA-Based Evolutionary Computation Model for Stock Market Forecasting

Stock market forecasting is used to draw attention of researcher since long and it will be. In this paper, a Data Envelop Analysis-based Gene Expression Programming model has been proposed and experimented with real data from BSE Sensex. The DEA has been used for filtering independent variables to be used as input variable of the GEP model. Different experiments have been made by first allowing all input variables to the GEP model directly without filtration by DEA and then allowing only those variables which are tested and marked as better variable to explain target variable. The result obtained from both the experiment has been put side by side and explained. From the analysis, it was noticed that the DEA-based GEP has better capabilities to forecast than the other one, even with less number of input variables.