Telecom-Churn-Reduction

Problem Statement :

Churn (loss of customers to competition) is a problrm for companies because it is more expensive to acquire a new customer than to keep your existing one from leaving. This problem statement is targeted at enabling churn reduction using analytics concepts.

Our task is to build classification models which will tell us whether a customer will churn or not. This repository contains train and test data set which we are using to predict the model to churn or not.

Model Selection :

From the train data we can say that the dependent variable is categorical that is true or false , after the EDA we have conclude to go with the Ensemble model that is RandomForest(RF) Classifier with the best grid search method as there are more features and RF can help to build complex model over decision tree.