Classification and analysis of customer data using a novel criterion based random forest algorithm to improve retention rate over SVM algorithm in terms of prediction rate

The framework aims at analyzing the telecom customer data and predicting the churn for improving the customer retention rate. The customer information dataset used for training and testing of the proposed prediction model consists of 7043 customers with 21 attributes. The Prediction of Churn is done by adapting Random Forest (RF) and Support Vector Machine (SVM) algorithms. The classification accuracy of the Random Forest classifier is (79%) and SVM is (75%). There is a statistically significant difference among the study groups with a significance value (p