摘要：In the establishment of a personal credit assessment model the prediction accuracy is very important. For the deficiency of the single model in personal credit assessment, this study puts forward the method of using the combined forecasting model to carry on it. Based on the advantages of different single model in personal credit assessment, the logistic regression and linear regression are chosen to construct this combination model and the improved cuckoo search algorithm (CS) is used to solve the model weights, a combination forecasting model based on CS is constructed for personal credit assessment. The results show that the model construct in this study can effectively reduce the second class misjudgment rate of personal credit assessment and effectively improve the forecasting accuracy, which has better applicability and significant value for commercial banks to control the consumer credit risk.
