Ensemble decision trees Maciej Zieba

This book concentrates on the problem of credit assignment. The goal of this work is to propose ensemble classification method, which can be applied to solve credit decisions problems. The method uses switching class labels techniques to construct complex classification system, which solves two of typical data mining problems observed in financial data mining field: the problem of imbalanced data and the problem of asymmetric cost matrix. The performance of the proposed solution is going to be evaluated on German Credits dataset. Это и многое другое вы найдете в книге Ensemble decision trees (Maciej Zięba)