Forecasting Customer Attrition using Machine Learning

Customer churning has various problems in a number of different industries. Due to the loss of customers, the companies not only have a lower revenue but there is a vast effect on the reputation and client number. Therefore, it has become crucial for businesses to anticipate customer churn. In this work, a model is created for predicting customer churn using artificially generated data. The work involves creating artificial data that closely resembles the traits of actual customer data. To create a dataset that accurately reflects the customer data, a framework for the generation of synthetic data is developed. To make sure the synthetic data is appropriate for machine learning algorithms, it is pre-processed and given engineered features. Using the simulated data, the assessment on how well different machine learning algorithms predicts the customer churn is estimated. The model's performance will be contrasted with that of conventional models developed using real world data. Later the maximum accurate machine learning model is deployed into a web application using flask for companies to input data and predict customer churn.