Institutional Repository of Vilnius University
Personalizuotas rinkodaros modeliavimas /
Abstract
dc:descriptionIn the modern world, as competition increases, the importance of cost reduction also grows. Many studies have found that the cost of attracting new customers is up to ten times higher than maintaining existing ones. For those reasons, the importance of customer loyalty analysis is increasing. Rapid technological development and their application in everyday life have already affected business models of many companies. Businesses integrate machine learning into their internal processes for a variety of strategic reasons. One of such reasons is the ability of the algorithm to adapt to data, recognize correlations and provide useful insights that could improve client segmentation. This would allow for more effective planning and implementation of marketing campaigns. The aim of this research was to assess and predict customer churn on telecommunications data and to create the most suitable model for determining this risk. In order to predict the churn, four models were applied: Logistic Regression, Decision Tree, Random Forest and XGBoost - gradient-boosted decision tree. The results concluded that the best model for predicting customer churn is XGBoost. Model’s ROC graph showed that the goodness on the test set reached 94.15%. The second best model for predicting customer churn is the Random Forest, with an accuracy on the test set of 93.65%.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ramanauskas, Tomas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192845471&prefLang=en_US