| Title: | Comparative analysis of machine learning models for telecommunications churn prediction |
|---|
| Authors: | ID Cerjan, Maja (Author)
ID Mršić, Leo (Author)
ID Rabuzin, Kornelije (Author)
ID Boshkoska, Biljana Mileva (Author) |
|---|
| Files: |  RAZ_Cerjan_Maja_2025.pdf (12,52 MB)
MD5: 2EBEEB33F12E23AE224DC803CB606675
|
|---|
| Language: | English |
|---|
| Work type: | Unknown |
|---|
| Typology: | 1.08 - Published Scientific Conference Contribution |
|---|
| Organization: | FIŠ - Faculty of Information Studies in Novo mesto
|
|---|
| Abstract: | Customer retention is a major problem in the telecommunications industry. This study develops and evaluates models to identify possible churners. Machine learning techniques (“Decision Trees”, “Random Forests”, “Logistic Regression” and “Neural Networks (multilayer perceptron MLP)”) were applied through Python and R to analyze the “Telco Customer Churn” Kaggle dataset, based on customer assests and service usage. The data pre-processing compiled missing data and then standardized it. Evaluation used nested 10-fold cross-validation with an inner loop for hyperparameter tuning and mutual-information top-K feature pruning, with pre-processing confined to training folds. In Python, RF and LR achieve F1(~0.629), with Logistic Regression accuracy ~0.75. In R, Logistic Regression performed best (F1 ≈ 0.60 ± 0.03, Accuracy ≈ 0.80 ± 0.01). Metrics derived from pooled confusion matrices averaged over folds equal outer-fold means, confirming generalization across folds and between Python and R. Research offers empirical evidence for transferring and testing churn prediction models across Python and R in telecommunications analytics, with fully reproducible evaluation and results. |
|---|
| Keywords: | customer churn, telecommunications, churn prediction, logistic regression, neural networks (MLP), Python, R, nested cross-validation |
|---|
| Publication status: | Published |
|---|
| Publication version: | Version of Record |
|---|
| Year of publishing: | 2025 |
|---|
| Number of pages: | Str. [133-146] |
|---|
| PID: | 20.500.12556/ReVIS-13034  |
|---|
| COBISS.SI-ID: | 265024515 |
|---|
| UDC: | 004.85:654 |
|---|
| Note: | Nasl. z nasl. zaslona;
Opis vira z dne 16. 1. 2026;
Soavtorji: Leo Mršić, Kornelije Rabuzin, Biljana Mileva Boshkoska;
|
|---|
| Publication date in ReVIS: | 22.01.2026 |
|---|
| Views: | 113 |
|---|
| Downloads: | 1 |
|---|
| Metadata: |    |
|---|
|
:
|
Copy citation |
|---|
| | | | Share: |  |
|---|
Hover the mouse pointer over a document title to show the abstract or click
on the title to get all document metadata. |
