Customer Churn prediction in Software as a Service (SaaS) Industry using Machine Learning Algorithms

The sole reason for an organization to exist in this era of competition is to remain profitable andcompetitive within the industry. There is no doubt in the notion that customers directly influence theprofitability of a firm, the customer can be categorized as existing and new customers. Although primarymotive of a firm is to cater more percentage of target customers, the ability of a firm to retain their existingcustomers is equally important. Churn rate can be defined as the rate at which the existing customers thatwere using a service or product had stopped using it. This study is focused on predicting the churn rate ofcustomers using various machine learning algorithms including Voting Classifier, Gradient BoostClassifier, AdaBoost, Logistic Regression and Gaussian Naïve Bayes Algorithm in the Software as aService (SaaS) industry on the basis of features selection from the collected data. The algorithms will beimplemented to identify the best algorithm with the most accuracy as compared to other algorithms. Thispaper contains one extra algorithm of Voting Classifier which shows highest accuracy with one moreevaluation matrix feature of F2-score for every algorithm implemented in the paper. This alsoimplemented two special classifiers for feature prediction giving different results using Gradient Boost andAdaboost classifier.