Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Jurnal Komtika (Komputasi dan Informatika)
Illah, Ibnu Zahy' Atha
Unknown Affiliation
Computer Science & IT Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Implementasi Metode Klasifikasi LightGBM dan Analisis Survival dalam Memprediksi Pelanggan Churn Illah, Ibnu Zahy' Atha; Jauharis Sapu, Wahyu Syaifullah; Damaliana, Aviolla Terza
Jurnal Komtika (Komputasi dan Informatika) Vol 8 No 1 (2024)
Publisher : Universitas Muhammadiyah Magelang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31603/komtika.v8i1.11194

Abstract

Increasingly tight competition in the business world causes every business sector to try to utilize relevant technology to maintain its market share. The success of a company is often measured by how strong the customer network they have. Loss of customers (customer churn) can cause a significant decrease in revenue and can even threaten the existence of the company itself. Therefore, predictive modeling and projection of customer churn is needed as a customer retention effort. This research involves the LightGBM classification algorithm for customer churn prediction and utilizes survival analysis for future projections. The results of the research can be used to prevent customer churn at companies, especially PT Kasir Pintar Internasional. LightGBM classification performance as measured by model evaluation reaches Accuracy, Precision, Recall, and F1-score values of 0.964, 0.971, 0.990, and 0.980 respectively. The LightGBM classification model also provides information on five important features that influence customer churn. Companies can use these five important features as material for designing customer retention strategies. Apart from that, the Cox Proportional Hazard survival model has a C-index evaluation value of 0.83, which means it is quite capable of projecting customer survival. The survival model also shows that currently non-churn customers have an average survival expectation of 15 months.
Co-Authors Damaliana, Aviolla Terza Jauharis Sapu, Wahyu Syaifullah