On identifying potential direct marketing consumers using adaptive boosted support vector machine

Abstract

Identifying potential consumers for direct marketing to very large data is a difficult and impossible task to do manually. Therefore, the machine learning approach needs to help analyze the data to contribute in determining the marketing strategy policy. In this paper, vector machine support methods using the Adaboost algorithm are investigated to classify potential customers for direct marketing of large bank data marketing. The adaboost algorithm aims to build a better model than the model generated from a single classifier. Data obtained from UCI machine learning repository. Total data is processed as many as 9280 data with the number of attributes of 20 classes and 1 target class. Training data and test data are divided into 70% and 30%. This classification predicts the prospects for a deposit subscription. The results show that the SVM method using Adaboost algorithm obtained accuracy is 95.07% and the sensitivity is 91.65% higher than the ordinary SVM approach.