Predicción de rotación de clientes en la industria de las telecomunicaciones utilizando métodos de minería de datos

Resumen

At present, in competitive space between companies and organizations, customers churn is their most important challenge. When a customer becomes churn, organizations lose one of their most important assets, which can lead to financial losses and even bankruptcy.  Customer churn prediction using data mining techniques can alleviate these problems to some extent.  The aim of the present study is to provide a hybrid method based on Genetic Algorithm and Modular Neural Network to customer churn prediction in telecommunication industries and use Irancell data as a sample. The accuracy result of this study which is 95.5% get the highest accuracy rank in comparisons with the result of other methods, which shows using modular neural network with two modules of feedforward neural network and also using genetic algorithm to obtain optimal structure for modules of the neural network are the most important indicators of this method to each the highest accuracy result among the rest of methods.