A Collective Data Mining Approach to Predict Customer Behaviour

The customer churn is a demanding area of research in e-commerce. This introduces a number of challenges in discovering a more accurate prediction of the information. The classification algorithms are the supervised machine learning techniques. The various classification algorithm has its own advantages. In the proposed approach, the various classification approaches such as support vector machine, Bayesian Classifier and the Random Forest Algorithm are used to analyze the customer churn and a collective approach is applied to develop an effective model. The collective value from various classification methods makes the proposed method more effective. The accuracy metric, precision metric, recall metric and F-Score metric values are calculated to show the effectiveness of this approach. The feature selection method is implemented on the data. The method will lead the education of institutions to predict the churn status and help them to increase the customer.