Telecommunication customer churn prediction with predictive analytics | |
| Author | Vamshi, Puttakota |
| Call Number | AIT RSPR no.IM-22-04 |
| Subject(s) | Customer relations--Management--Data processing Business forecasting Machine learning Telecommunication--Customer services |
| Note | A research submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Information Management |
| Publisher | Asian Institute of Technology |
| Abstract | The telecommunications industry is one of the major sectors across the world and has many competitors of both government and private players. The telecom industry is a customer centric industry as it depends on its customers for the development of the company. The quality of the company depends on the number of customers of the company and the customer churn rate, so it is important for the companies to have customer relationship management and customer churn management as they are important parts of customer management. The customer churn rate is the percentage of customers who stop using the service or have not renewed their subscriptions of the company in a period of time. Having a lower churn rate than the competitors indicates that the company is in a better competitive state than its competitors, so it is important for companies to have a customer churn prediction system which helps them identify the customers who are likely to churn. The objective of this study is to compare different combinations of different sampling techniques and different prediction models and then find the best combination of the sampling technique and prediction model. The best combination found is then selected to be deployed using an application developed which helps in determining telecommunication customer churn prediction. In this study different sampling techniques such as Random Sampling, Random Undersampling, Near Miss Undersampling and Condensed Nearest Neighbor Rule for Undersampling are used to resample the data to handle the imbalance and remove the bias caused by the imbalance. The resampled data is given as the input to the different prediction models such as KNN, Neural network, Naïve Bayesian and Random Forest. The combination of sampling technique and prediction model with highest accuracy is then selected and deployed using an application which helps in determining whether the customers would churn from a telecom company or not. The application is built using flask, HTML, CSS and Bootstrap and the deploying of the prediction model is done using flask. |
| Year | 2022 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Information Management (IM) |
| Chairperson(s) | Vatcharapon Esichaikul |
| Examination Committee(s) | Chutiporn Anutariya;Huynh, Trung Luong |
| Scholarship Donor(s) | AIT Partial Scholarship |
| Degree | Research studies project report (M. Eng.) - Asian Institute of Technology, 2022 |