Performance Analysis of Hoeffding and Logisitic Algorithm for Churn Prediction in Telecom Sector

Churn automatically introduces where the term customer emanates. Churn is irresistible thus it is better to develop models to predict the churn. In this paper, weka tool is used to perform various operations on data set and first of all data is converted into nominal form. It's divided into classes and cross-validation is performed on churn class. Hoeffding tree algorithm and Logistic algorithm have been applied on the dataset. Classifier performance evaluator is used to classify the data. To observe the results text viewer is used and the graphical representation is cast-off to demonstrate the comparisons. Confusion matrix suggest that Logistic algorithm is a better algorithm to predict the customer churn over Hoeffding tree algorithm.