Financial Credit Default Forecast Based on Big Data Analysis

How to effectively evaluate and identify the potential default risk of borrowers calculate probability before issuing loans is basis important link credit management modern financial institutions. This paper mainly studies statistical analysis historical loan data banks other institutions with help idea non-balanced classification, uses machine learning algorithms (not algorithms) such as random forest, logical regression decision tree establish prediction model. The experimental results show that neural network forest algorithm outperform logistic classification in performance. In addition, by using rank importance features, features have a greater impact on final can be obtained, so make more effective judgment field.