The Research of Stock Predictive Model Based on the Combination of CART and DBSCAN

Abstract

Along with the development of electronic and intelligence in the world's stock market advances, the accumulation of the stock data grows larger over time. It is of great concern on the ways to find the hidden rules of information in the mass of data. Given the background above, this paper explores the methods of data mining by using the combination of Decision tree algorithm and Clustering algorithm. In addition, this paper accomplishes stock forecasting by combining CART algorithm and DBSCAN algorithm to build a predictive model with good applicability through a large number of experiments for parameter testing. According to the works above, the predictive model has a high accuracy and provides a scientific theory supporting the investment decisions.