Development of Churn Prediction Model using XGBoost - Telecommunication Industry in Sri Lanka

Prasanth, Senthan; Rathnayaka, R. M. K. T.; Kuhaneswaran, Banujan; Kumara, Banage T. G. S.

doi:10.1109/iemtronics52119.2021.9422657

Cited by 8 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance results during the forecasting process were assessed using several methods, including Accuracy, Confusion Matrix, Precision, Recall, and F1-Score, with Kfold cross-validation as the primary technique. [24]. Based on the results of the tests carried out, the values of each evaluation model can be seen in Table 2.…”

Section: Results Of Identifying Highly Correlated Featuresmentioning

confidence: 99%

Analysis Implementation of the Ensemble Algorithm in Predicting Customer Churn in Telco Data: A Comparative Study

Sari

Febriyanto

Adi

2023

IJCAI

View full text Add to dashboard Cite

Globalization and technological advancements in the telecommunication industry have led to a significant rise in the number of operators, leading to intense market competition. This sector has become crucial in developed countries, and companies strive to increase profits by acquiring new customers, up-selling existing ones, and extending the retention period of current clients. In the traditional method of defect prediction, a single classifier is used to build a model on a pre-labeled dataset. However, this approach has limitations in predicting defects accurately under certain circumstances. To overcome these limitations, boosting is applied to combine multiple weak classifiers and create a robust classification model. Among many algorithms used for churn prediction, ensemble techniques have demonstrated greater accuracy than simpler approaches. This study aims to overcome these limitations by experimenting with five ensemble algorithms, including Adaboost, Gradient Boost, XGBoost, CatBoost, and LightGBM. The results indicate that XGBoost outperforms other techniques and is the most suitable algorithm to build the predictive model. Additionally, the study achieves higher accuracy by performing a Grid Search CV hyper-parameter setting with XGBoost, resulting in an accuracy of 81.2%.Povzetek: Študija je primerjala pet ansambelskih algoritmov za napovedovanje prekinitve naročniškega razmerja. Rezultati kažejo, da je XGBoost najboljši algoritem z natančnostjo 81,2 %.

show abstract

Section: Results Of Identifying Highly Correlated Featuresmentioning

confidence: 99%

Analysis Implementation of the Ensemble Algorithm in Predicting Customer Churn in Telco Data: A Comparative Study

Sari

Febriyanto

Adi

2023

IJCAI

View full text Add to dashboard Cite

show abstract

“…Most of the use of the XGBoost algorithm to date has been used to develop customer churn prediction models. In the paper [36], [37], these two studies discuss the challenges of unbalanced data sets in the telecommunications industry and the variations in real telecommunications data compared to publicly available data sets. By utilizing the application of XGBoost Algorithm on this dataset, it achieves 97% of accuracy evaluation performance result and 88% of F1 score.…”

Section: F Xgboostmentioning

confidence: 99%

Comparative Analysis of Machine Learning Algorithms for 5G Coverage Prediction: Identification of Dominant Feature Parameters and Prediction Accuracy

Yuliana,

Iskandar,

Hendrawan

2024

IEEE Access

View full text Add to dashboard Cite

5G technology is a key factor in delivering faster and more reliable wireless connectivity. One crucial aspect in 5G network planning is coverage prediction, which enables network providers to optimize infrastructure deployment and deliver high-quality services to customers. This study conducts a comprehensive analysis of machine learning algorithms for 5G coverage prediction, focusing on dominant feature parameters and accuracy. Notably, the Random Forest algorithm demonstrates superior performance with an RMSE of 1.14 dB, MAE of 0.12, and R2 of 0.97. The CNN model, the standout among deep learning algorithms, achieves an RMSE of 0.289, MAE of 0.289, and R2 of 0.78, showcasing high accuracy in 5G coverage prediction. Random Forest models exhibit near-perfect metrics with 98.4% accuracy, precision, recall, and F1-score. Although CNN outperforms other deep learning models, it slightly trails Random Forest in performance. The research highlights that the final Random Forest and CNN models outperform other models and surpass those developed in previous studies. Notably, 2D Distance Tx Rx emerges as the most dominant feature parameter across all algorithms, significantly influencing 5G coverage prediction. The inclusion of horizontal and vertical distances further improves prediction results, surpassing previous studies. The study underscores the relevance of machine learning and deep learning algorithms in predicting 5G coverage and recommends their use in network development and optimization. In conclusion, while the Random Forest algorithm stands out as the optimal choice for 5G coverage prediction, deep learning algorithms, particularly CNN, offer viable alternatives, especially for spatial data derived from satellite images. These accurate predictions facilitate efficient resource allocation by network providers, ensuring high-quality services in the rapidly evolving landscape of 5G technology. A profound understanding of coverage prediction remains pivotal for successful network planning and reliable service provision in the 5G era.

show abstract

Comparison of Principal Component Analysis and Recursive Feature Elimination with Cross-Validation Feature Selection Algorithms for Customer Churn Prediction

Matin

Triayudi

Aldisa

2023

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

Development of Churn Prediction Model using XGBoost - Telecommunication Industry in Sri Lanka

Cited by 8 publications

References 12 publications

Analysis Implementation of the Ensemble Algorithm in Predicting Customer Churn in Telco Data: A Comparative Study

Analysis Implementation of the Ensemble Algorithm in Predicting Customer Churn in Telco Data: A Comparative Study

Comparative Analysis of Machine Learning Algorithms for 5G Coverage Prediction: Identification of Dominant Feature Parameters and Prediction Accuracy

Comparison of Principal Component Analysis and Recursive Feature Elimination with Cross-Validation Feature Selection Algorithms for Customer Churn Prediction

Contact Info

Product

Resources

About