Application of Random Forest on Selecting Evaluation Index System for Enterprise Credit Assessment
Peng Guo-lan · 2007
One of the key steps for the modeling of enterprise credit assessment is the selection of its evaluation index system.It affects directly the prediction accuracy of the model.A new ensemble-learning algorithm-Random Forest is applied in this paper.As a consequence,the selected index system is more objective and more suitable for machine learning.Numerical experiments show that the index system selected by Random Forest can effectively reflect the credit status of the enterprises,and improve the prediction accuracy of the assessment model based on Random Forest.