GA-Based Classification and Bank Loans Risk Early Warning

A new GA-based classification algorithm for bank loans risk early warning system is proposed in this paper. Training samples are supervised clustering by attribute similarity and class tags. The number and center of class families are automatically found by the fitness function as the goal constructed by the training samples. In this way, it can guarantee that they tend to the global optimum under the influence of subjective factors. A classification of the test sample is determined by the class tag of the class family and the nearest neighbor rule between the test sample and the center of class families. According to a domestic bank credit data, it is verified that the algorithm improves the accuracy of the early warning system by 13 percent in contrast to the existing result.