Application Study of Corporate Profit Prediction Using Decision Tree Ensemble Model

摘要

In this paper,we apply ensemble method to predict the future profit states of the listed enterprises.We take decision tree model as the basic classifier to construct the ensemble model by means of Bagging technique.The empirical result shows that the rate of prediction accuracy of the ensemble model is greater than 96%.We argue that the ensemble model can improve the prediction accuracy and is more robust to the individual decision tree model.By comparing the results derived from different ensemble models which are constructed by different number of base classifiers,we find that the ensemble model constructed by 35 base classifiers works best.