AI Based Nose for Trace of Churn in Assessment of Captive Customers

Main Article Content

Drumil Joshi , Aayush Parikh , Ritika Mangla , Fawzan Sayed , Dr. Sunil H Karamchandani


Customer churn speculation is aimed at finding customers with a high potential for attraction. Predictive accuracy, Precision, and justification are the three critical elements of a churn predictive model. According to domain knowledge, the Accurate standards of the model allow us to identify the main drivers for customers to churn and develop an effective retention strategy. In this research paper, we present a comparative study of the most popular machine learning classifiers used to solve the problem of churning customers in the telecommunications sector. In the first phase of our test, all models were implemented and tested using statistical evaluative measures on the popular telecom database. In the second phase, the performance improved by boosting was studied. In order to determine the most efficient parameter combinations, we performed hyperparameter tuning for the best classifier and a wide range of parameters. The best overall classifier was XG Boost classifier after Hyperparameter tuning with an accuracy of almost 82% and Precision of 0.8.

Article Details