Analysing Customer Churn and Customer Attitude in Telecom Market

In the telecom market, churn prediction models are developed by academicians to check and control the migration of the customer and to retain the existing customers. Extensive literature studies based on marketing concepts affirm that retaining the existing customers is more important and easier that to acquire a new customer. And so the market researchers are in a position to analyze and design a suitable model for predicting customer churn and evaluating customer attitude. However, in most cases, it turned out that most of the used techniques to solve this problem fail to address the complex relationship between customer features and churn. A novel mathematical model based on Neural Networks for predicting the customer churn and evaluating the customer attitude is proposed in this paper. The method used in the proposed model tries to find the pattern to identify the possible churners and their attitude based on the historical data. To predict the churning rates of the mobile phone users, two established Recurrent Neural Network algorithms, viz. Elman Recurrent Neural Network (ERNN) and Jordan Recurrent Neural Network (JRNN) are applied and their performance is evaluated. Both the algorithms are tested using the real time data information collected from the mobile phone users of the Indian Public Sector Telecommunication Company, Bharat Sanchar Nigam Limited (BSNL) and the detailed experimental study is made. From the analysis, it is found that the performance of JRNN is better when compared to ERNN. The Experimental analysis is tested using the Software tool, MATLAB. To reduce cost and maximize effectiveness, churn prediction has to be as accurate as possible to ensure that only the customers who are planning to switch their service providers are being targeted for retention. Our experimental results show that Recurrent Neural Networks serve the purpose with maximum accuracy.