Classification and Regression Trees and Their Use in Financial Modeling
暂无分享,去创建一个
Classification and regression trees (CART) are nonparametric and nonlinear modeling techniques that do not rely upon the many stringent assumptions required by classical parametric models. Despite the fact that researchers in many fields have regularly found trees to be an attractive way to express underlying relationships, they are relatively unfamiliar to financial modelers where the historical focus of financial modeling has been on parametric regression. Although the linear type of regression analysis is convenient and sometimes intuitive, it may not fully capture the complexity of financial markets. As the quantity and variety of financial information available to data exploration have increased over time, there has been a commensurate need for a more robust and versatile process to analyze these data. CART offers a valuable alternative to traditional methods for modeling financial data.
[1] E. Sorensen,et al. The Decision Tree Approach to Stock Selection , 2000 .
[2] H. Frydman,et al. Introducing Recursive Partitioning for Financial Classification: The Case of Financial Distress , 1985 .
[3] KaoDuen-Li,et al. Equity Style Timing (corrected) , 1999 .
[4] Leo Breiman,et al. Random Forests , 2001, Machine Learning.
[5] Min Zhu,et al. A Hybrid Approach to Combining CART and Logistic Regression for Stock Ranking , 2011, The Journal of Portfolio Management.