Study on call detail records of family members based on classification model

Abstract

In telecommunication industry, machine learning techniques have been applied to the Call Detail Records (CDRs) for predicting customer behavior. To further investigate the information buried in huge amounts of CDRs, family relationship among mobile users can be identified, which helps the effective targeted marketing behavior, it is significantly important for increasing profitability. We use the information extracted from the CDRs analysis to identify customer calling patterns. then Customers calling patterns are given to a classification algorithm to generate a classifier model for predicting the family relation of a customer. We apply different machine learning techniques to build classifier models and compare them in terms of classification accuracy and computational performance. The reported test results demonstrate the applicability and effectiveness of the proposed approach.