Predicting customer churn rate with machine learning, for use in customer lifetime value calculations in the telecommunications industry
(2024) EXTM10 20241Department of Economics
- Abstract
- In the rapidly evolving telecommunications industry, customer retention is crucial for maintaining profitability. This report investigates the performance of various machine learning models in predicting customer churn rates, which are integral to calculating Customer Lifetime Value (CLV). The focus is on a comparative analysis of several models including logistic regression, ridge regression, lasso regression, random forest, AdaBoost, XGBoost, and support vector machines, using two different datasets from the telecom sector. Our research explores the generalizability of churn prediction models, aiming to identify whether a particular model consistently excels across various datasets. Additionally, the study examines the appropriate... (More)
- In the rapidly evolving telecommunications industry, customer retention is crucial for maintaining profitability. This report investigates the performance of various machine learning models in predicting customer churn rates, which are integral to calculating Customer Lifetime Value (CLV). The focus is on a comparative analysis of several models including logistic regression, ridge regression, lasso regression, random forest, AdaBoost, XGBoost, and support vector machines, using two different datasets from the telecom sector. Our research explores the generalizability of churn prediction models, aiming to identify whether a particular model consistently excels across various datasets. Additionally, the study examines the appropriate evaluation metrics for assessing model performance, highlighting the significance of the Brier score in calculating churn probabilities. The findings indicate that XGBoost is the top-performing model on both datasets, leading to the recommendation for telecom operators to adopt XGBoost for their churn rate calculations. However, since model effectiveness varies by dataset, model selection should be done with caution. Selecting a model should consider more than just performance, including factors like interpretability and ease of implementation. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9170997
- author
- Karnehed, Joel LU and Regnell, Albert LU
- supervisor
- organization
- course
- EXTM10 20241
- year
- 2024
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- Churn Rate Analysis, Customer Lifetime Value, CLV, Logistic regression, Machine Learning, CART, Data Imbalance
- language
- English
- id
- 9170997
- date added to LUP
- 2024-11-22 09:05:14
- date last changed
- 2024-11-22 09:05:14
@misc{9170997, abstract = {{In the rapidly evolving telecommunications industry, customer retention is crucial for maintaining profitability. This report investigates the performance of various machine learning models in predicting customer churn rates, which are integral to calculating Customer Lifetime Value (CLV). The focus is on a comparative analysis of several models including logistic regression, ridge regression, lasso regression, random forest, AdaBoost, XGBoost, and support vector machines, using two different datasets from the telecom sector. Our research explores the generalizability of churn prediction models, aiming to identify whether a particular model consistently excels across various datasets. Additionally, the study examines the appropriate evaluation metrics for assessing model performance, highlighting the significance of the Brier score in calculating churn probabilities. The findings indicate that XGBoost is the top-performing model on both datasets, leading to the recommendation for telecom operators to adopt XGBoost for their churn rate calculations. However, since model effectiveness varies by dataset, model selection should be done with caution. Selecting a model should consider more than just performance, including factors like interpretability and ease of implementation.}}, author = {{Karnehed, Joel and Regnell, Albert}}, language = {{eng}}, note = {{Student Paper}}, title = {{Predicting customer churn rate with machine learning, for use in customer lifetime value calculations in the telecommunications industry}}, year = {{2024}}, }