Author: Swathy K,Shalini S,Selvakumar K,Kanagaraj S
Date Published: 03 Jan 2023
The concept of Churn has been used for years to increase profitability and stabilize customers-company relationships. So this is becomes the major issue in business industry in multiple domains such as Retail, Telecom industry, Finance, Online Music Streaming, Online Game, Internet Funds, Insurance etc. The motive is to predict whether the users will churn or not. There are some hypothesis to find whether the activities of the customer. In this paper we are using Sparkify dataset to predict the churn activities. Sparkify is a Online music Streaming application which is similar to Spotify. In this paper we have compared the important features using Particle Swarm Optimisation(PSO). Here we have compared many algorithms like, Deep Neural Network from Deep Learning, Light Gradient Boosting, Bagging from Machine Learning Algorithms along with some of the metrics like Senstivity, Specificity, Accuracy, Precision etc.