Using data mining techniques to predict the success of bank telemarketing

Classification model is a main data mining technique in bank prediction. The purpose of this paper is predict the success of telemarketing deposit. The data were collected from marketing campaigns of a bank institution in Portuguese country. The marketing was relying on telephone calls. The Bank additional Dataset was randomly chosen from 4119 examples, out of 41188 with 20 features, collected data from 2008 to 2010. Moreover, five DM models such as Naive Bayes (NB), Logistic Regression (LR), decision tree (DT), Multilayer perceptron Neural Network (MLPNN), and Support vector machine (SVM) were compared. Four metrics used as receiver operating characteristic (ROC), Accuracy (ACC), sensitivity and specificity to find the better the results. The LR presented the best results with ROC = 0.93 and ACC = 91.21%. These results were confirmed the LR model is the best choice for the success of bank telemarketing campaign managers.

کلیدواژه ها:

Direct Marketing ، Classification ، Naive Bayes ، Multilayer Perceptron neural network ، Logistic Regression ، Data Mining ، Telemarketing