Prediction of choosing types of insurance for the customers

Abstract

The objectives of this research are to study the factors that have an influence on the decision making for choosing types of insurance of customers and to compare the efficiency of four classification methods which are multinomial logistic regression, discriminant analysis, decision tree and artificial neural network. By considering of six independent variables, namely gender, age, monthly income, education, marital and occupation. This study is found that multinomial logistic regression analysis indicates that five factors affecting the decision making are age, monthly income, education, marital, and occupation. In addition, the study of discriminant analysis indicates that three factors; gender, age and marital are significantly affecting the decision making. For the efficiency comparison of four methods, it’s shown that neural network is the most efficiency method which have the accuracy is 85.57% and followed by decision tree, multinomial logistic regression analysis, and discriminant analysis which have the accuracy are 76.50%, 70.00% and 61.80%, respectively.