Ensemble classification method for daily return stock market

Abstract

Stock price movements in Indonesia are measured using indices, one of which is the Composite Stock Price Index (CSPI). CSPI is a stock index that measures a combination of all shares from various sectors listed on the Indonesia Stock Exchange. The ensemble method is to build predictive models by combining the strengths of the classical classification method. In this research, the purpose of ensemble based on Boosting for Regression appeared to enhance simple tree analysis and deals with some of the weaknesses found in uncomplicated techniques. The ensemble tree combines the prediction values of many simple trees into a single prediction value. Based on the experiments that have been carried out, the ensemble method proved to have a better accuracy rate, which amounted to 82%. It is assumed that the ensemble model can obtain the relationships between variables that are more precise than the previous model.

Keywords

• ensemble method
• stock market returns
• classification