After using data mining method to analyze, the improvement of judgment efficiency of the problem (customer identification, etc. ).

For example, the default rate of loan customers is 2%. You use data mining to build a model to predict the possibility of loan customers' default in the future. After scoring customers with this model, among the 10% customers with the highest possibility of default, the actual default customers account for 50%, which means that the recognition rate of your model for this part of customers is 25(50% is 25 times that of 2%), so you just need to do it well.