Machine Learning and Deep Learning technique used in Customer Churn Prediction: - A Review

Churn Prediction is crucial in a number of service- based businesses, including the telecom sector, life insurance, hospitality, banking, and gaming. A Prediction of consumer churn in the enterprises is crucial because it is one of the major risks to revenue loss and one of the top concerns for big enterprises. As per studies it is found that the expense of acquiring a new consumer is 10 times greater than the expense of an existing consumer. This issue has an impact on both the company's revenue and its ability to grow. One of the most important tasks for the business is to keep its current customers. As technology advances quickly, Deep Learning and Machine Learning techniques are evolved that businesses can employ to track customer churn behaviour. In this review we will cover, overview on Customer churn, Machine learning Classification and boosting Models as naive bayes, support vector machines, k-nearest neighbor, logistic regression, decision trees, Random Forest, Adaboost, XGBoost and artificial neural networks algorithms, as well as deep learning techniques. We will also explain performance metrics like confusion matrix, Recall, precision, F-measure which is utilized in the churn prediction models. As accuracy is a decent performance indicator, but assessing performance solely by accuracy is insufficient because accuracy on small datasets is more consistent and predictable. In addition to accuracy, researchers must also consider other performance.