International Research Journal of Engineering and Technology (IRJET) Volume: 11 Issue: 04 | Apr 2024
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
Churn Prediction Modelling Using Regression Techniques Akansha Shukla1, Rakesh Kumar2 1PG Student, Dept. Of Computer Science Engineering, Madan Mohan Malviya University of Technology, Uttar Pradesh,
India 2Professor, Dept. Of Computer Science Engineering, Madan Mohan Malviya University of Technology, Uttar Pradesh,
India ---------------------------------------------------------------------------***-----------------------------------------------------------------------ABSTRACT — An important application in the study of customer behavior is the prediction of customer churn or the probability that a client will move to a rival. It is usually less expensive to keep current clients than to find new ones. Although it can be difficult, predicting consumer behavior is essential for service-based firms. In this paper, precise forecasts are generated by the utilization of data mining tools. In the banking sector, customer attrition happens when customers stop using the products and services the bank provides for a while and then cut off communication with the bank. In light of this, maintaining customers is crucial in the fiercely competitive banking industry of today. The basis for forecasting future clients, and the source of churn is past data. A statistical model has been built to predict the response for current customers by looking at the data of customers who have already churned (response) and their traits/behavior (predictors) before the churn occurrence. This strategy is classified as supervised learning. This study uses a large-scale, unbalanced dataset from a bank to forecast client attrition using a logistic regression model. In terms of predicting customer turnover, this method's performance was compared to that of the decision tree, K-nearest neighbor, and random forest classification models. This task aims to suggest the approach that yields the best accuracy rate, recall, and precision scores. These metrics are useful in gauging the bank's capacity to predict client attrition.
model that can determine the likelihood of client churn. Our main objective is to identify the key factors that can accurately predict the churn rate among customers. The role of a predictive model is to bring the churned customers to light. The proposed model's purpose is to bring churned customers to light. In a targeted approach industry tries to identify which customers are likely to churn. The industry then targets those customers or clients and provides them with special incentives, offerings, and plans except for normal customers. This approach can bring a huge loss to the industry, if churned measures are inaccurate because the industries are wasting a lot of money on the customers who would have stayed anyways, irrespective of short or long distance. It's being used in every field [2-4]. To achieve this, we examine a comprehensive dataset that includes information such as customer credentials, gender, dependents, city, branch code days since the last transaction, and occupation, among other variables. To narrow down our analysis, we specifically focus on the occupation variable and divide it into subcategories. Through our analysis, we explore the interactions between these variables and the customer's balance, ultimately enabling us to predict churn for new candidates based on their credentials. By applying our models to the training data, we can predict the dependent variables for the test data. Subsequently, we examine our solution to identify the features that have the greatest impact on predicting churn.
Keywords — Churn prediction, Machine Learning, Logistic Regression Modelling, Supervised Learning
1. INTRODUCTION Customer attrition, also known as customer churn, is the phenomenon where customers terminate their relationship with a business or organization. In the context of banking, customer attrition occurs when customers close their accounts or discontinue utilizing the service of a particular bank.[1] By implementing strategies for churn prevention, companies can develop loyalty programs and retention campaigns to retain as many customers as possible. In this particular project, we utilize customer data from a banking institution to construct a predictive
© 2024, IRJET
|
Impact Factor value: 8.226
Figure 1:- Workflow Diagram Banking is one of the sectors where analyzing customer behavior and estimating customer churn based on these behaviors is an essential topic of research. Customer churn analysis results have a large impact on the bank's policy.
|
ISO 9001:2008 Certified Journal
|
Page 2036