Building a cash advance assessment model based on machine learning approach

Credit risk is an important type of risk in the financial field, and its management is aimed at management aimed at identifying, assessing and controlling this risk. In order to establish a reasonable credit assessment system can effectively manage and control credit risk, this paper selects desensitized customer data from the data warehouse according to the ER relationship graph, the size of the data set is 100,000 items, a total of 10 features, by filtering the features and using the machine learning method to predict the labels of the data set, and then calculate the AUC as well as the KS value. In this paper, it is found that the lightgbm model performs significantly better than other models in out-of-time samples for prediction, which has the value of scorecard formulation for this dataset. Ultimately, lightgbm scorecard is created based on the prediction results and target users are filtered based on their scores.