The Application Of Decision Tree In Bank Telemarketing

At present, with the development of the Internet Banking, the application of Internet Technology in the financial industry is increasingly. As a new data analysis technology, Data Mining provides a possibility for the analysis and processing of a large number of historical data. Therefore, in this paper, the decision tree classification algorithm is applied to the classification of bank customers, due to the low efficiency and accuracy of traditional decision tree, we further propose the ensemble methods.First of all, this paper describle the theoretical basis of Data Mining and classification techniques, and gives the basic principles and different of te ID3, C4.5, CART algorithm, and gives advantages and disadvantages of several basic decision tree algorithms.Secondly, in this thesis, the basic decision tree algorithm is applied to bank deposit data set. Feature selection selects “job, marital, education, duration, poutcome” as the optimal variables, the decision tree classification model gives eight classification rules, the very useful information can be found from the classification rules. And they can provide a scientific basis and accurate decisions for the Financial practitioners.Finally this paper presents AdaBoost lifting algorithm and the Random Forest algorithm bacause of the basic decision tree algorithm's fitting effect is not ideal. The imporved decision tree algorithm is applied to the data and it proved that the imporved decision tree algorithm's classification effect is significantly better than the basic decision tree classification algorithm.The decision tree classification model established in this paper fits the bank data set prefectly. The classification rules can be very helpful for the financial institutions in the business marketing to reduce marketing expenses, access to high value customers, it provides a reference for the scientific decision-making of the relevant departments in financial industry.