Back to search

Institutional Repository of Vilnius University

Churn prediction in telecommunication industry using machine learning /

Abstract

dc:description

The digitization and rapid growth of technological advancement has changed the way companies do business. With more and more services to chose from, consumer churning has become one of the biggest threats to all companies. In this paper we analyze a machine learning based churn prediction model for a telecommunication service provider. The main aim is to find the best churn prediction model and implement the model on real data. Three machine learning algorithms are compared: Random Forest classifier, XGBoost classifier and Support Vector Machine classifier. The models are evaluated using the metrics, accuracy, precision, recall, and F-scores. The prediction performance is tested on a data set provided by one of the telecommunication companies in Lithuania. It was found that the data is imbalanced with a majority of non-churners, so we also used two balancing techniques, Random Under-Sampling and Random Over-Sampling, to see if they improve the results of our models. We conclude that machine learning is a very useful process for customer churn prediction and that Random Forest and XGBoost models are better at predicting churn than the Support Vector Machine model.

Degree

thesis:*
Grantor dc:publisher
Institutional Repository of Vilnius University
Year dc:date
2022

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Latvaitis, Tautvydas,

Subjects

dc:subject × 1

Rights

dc:rights
Statement dc:rights
  • info:eu-repo/semantics/openAccess
Language dc:language
eng

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
Vilnius University
Base URL
epublications.vu.lt/oai
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Latvaitis, Tautvydas,. Churn prediction in telecommunication industry using machine learning /. Institutional Repository of Vilnius University, 2022. https://repository.vu.lt/VU:ELABAETD146235608&prefLang=en_US