Back propagation neural network approach for churn prediction: a case study in Celcom (M) Berhad / Hapida Husin

One of the main objectives of modeling customer churn is to determine the causal factors, so that the company can try to prevent the attrition from happening in the future. Churn prediction for mobile telecoms companies is quite complicated but chum is very important as it is a measure of customer loyalty, and therefore how stable a company's subscription revenues are likely to be if sales growth flags. This report presents the development of an application in the Artificial Neural Network (ANN). The ANN is utilised to predict the churn from a set of historical customer data from Celcom (M) Bhd. The data has been chunk into a few series to determine the significant variables for predicting churn. Computer program were written in Matlab to implement the training and testing programs for the ANN algorithms. At the end of the study, results showed that the developed technique is capable and feasible to be used in practical. The developed ANN can be utilised for predicting churn and the system stability can be evaluated using new set of data.