The Application Of Data Mining In Bank Telemarketing

With economical globalization and liberalization, the banking industry is confronted with all-around and multilevel market competitions. The rapid development of the Internet age driven by technology turns the traditional passive service to active customer care. In the field of marketing, the traditional extensive customer marketing strategy is becoming more and more popular than the intensification of customer marketing strategy.Taking customers as the center of marketing activities is becoming a new favorite strategy.Taking predicting the bank telemarketing results as the research object, firstly, the paper will introduce the research problem mainly in three aspects: the background? the current research status at home and abroad and the research methods. Secondly, the paper will give a brief introduction about the bank telemarketing in big data era,including bank telemarketing relevant theories, the BP neural network, classification of decision tree?the support vector machine(SVM) method and marketing in big data era.Neural network method is a parallel distributed processing model. By training support vector machine(SVM) classifier can we get the minimum errors of small samples,having a very good classification and generalization ability. Classification rules of decision tree is established on a variable value separation. The classification of the decision tree is a tree shaping structure model.Under the background in big data era, when integrating application of neural network,decision tree and support vector machine(SVM) classification of data mining method to establish the bank telemarketing model, and using case analysis and comparative analysis,we found that the best result is the BP neural network, and the AUC value is 0.97 according to the ROC curve, response rate curve and capture rate curve. Suggestions put forward on the telephone marketing are as following:changing the bank customer selection strategy, choosing the most possible customers, reducing the cost of customer contact,improving efficiency, and creating more value. In the future, sales managers can getvaluable information by making full use of highly correlated properties. We hope that such suggestions can be provided for domestic Banks.