The Decision Tree Model In Bank Risk Early Warning System

With China's earning of the membership of the WTO,the competition within banks is increasingly severely.The bank earlywarning system is one important branch of the bank risk system, and plays an important role in inhencing the ability of competition for the banks.In current system,the classification of customers, subsidiary bank and new business still stays in the phrase of manual working.To reduce the bad influnce of human factors,and acquire more accurate classfication results,this disseration will analyze the feasibility of applying decision tree in the bank earlywarning system,then it will propose a new executive plan, which aims at increasing the quality of earlywarning,and this plan will apply the in the bank earlywarning system.Classification is an important part of data mining,which establishes a model relying on data of known class and then forsees the class of the unkown-class data with the model.Common methods for classification are decision tree,Bayesian,neural net and so on;among these methods the method outperforms other methods in easy to understand,facility for training,implementation and utility.This dissertation help me choose decision tree as my classification method.According the executive plan,this disseration will examplify to realize the whole process of decision tree completely, it contains:ensure the object; get needed data;use the data pretreatment method such as data data integration,data clean,data transformation and so on;use ID3 algorith to make the decision tree;get the classification rule according to the tree to complete the decision tree model;Finally,the decision tree model aris successfully applied into the bank earlywarning system and gains an upstanding effect.