Muhamad Ilyas Haikal, . (2024) PENGEMBANGAN MODEL PREDIKSI CHURN PADA INDUSTRI TELEKOMUNIKASI DENGAN METODE LOGISTIC REGRESSION. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.
![[img]](http://repository.upnvj.ac.id/style/images/fileicons/text.png) |
Text
AWAL.pdf Download (988kB) |
|
Text
ABSTRAK.pdf Download (144kB) |
|
Text
BAB 1.pdf Restricted to Repository UPNVJ Only Download (199kB) |
|
Text
BAB 2.pdf Restricted to Repository UPNVJ Only Download (453kB) |
|
Text
BAB 3.pdf Restricted to Repository UPNVJ Only Download (533kB) |
|
Text
BAB 4.pdf Restricted to Repository UPNVJ Only Download (852kB) |
|
Text
BAB 5.pdf Download (176kB) |
|
Text
DAFTAR PUSTAKA.pdf Download (178kB) |
|
Text
LAMPIRAN.pdf Restricted to Repository staff only Download (607kB) |
|
Text
RIWAYAT HIDUP.pdf Restricted to Repository staff only Download (102kB) |
|
Text
HASIL PRAGIARISME.pdf Restricted to Repository staff only Download (10MB) |
|
Text
ARTIKEL KI.pdf Restricted to Repository staff only Download (120kB) |
Abstract
With the growing number of internet and telecommunication service users in Indonesia, the number of telecommunication companies is also increasing. This growth has led to a higher incidence of Churn, where Customers terminate their telecommunication services. Companies prefer to retain Customers as it is less costly than acquiring new ones. However, current strategies to reduce Churn are still facing various challenges. This study aims to create an accurate Churn Prediction model and develop effective strategies to reduce Churn. This research employs various Prediction models, focusing on Logistic regression with the (RSCV) Random Search Cross Validation and SMOTE (Synthetic Minority Over-sampling Technique) method. The LRG SMOTE model stands out with an accuracy of 74.2% and the highest recall of 0.820, demonstrating its effectiveness in identifying Customers who are likely to Churn. Coefficient analysis indicates that features such as long-term contracts and telephone services significantly reduce Churn, while higher monthly charges increase the risk of Churn. Further, Recursive Feature Elimination (RFE) analysis identifies Total Charges, Monthly Charges, and Tenure as the main predictors of Churn. High correlations between security services like OnlineSecurity and OnlineBackup suggest opportunities for cross-selling strategies, while potential bundling strategies are indicated by correlations between StreamingTV and StreamingMovies.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Additional Information: | [No Panggil: 2010314056] [Pembimbing: Fajar Rahayu,.] [Penguji 1: Ferdiyanto, .] [Penguji 2: Silvia Anggraeni,.] |
| Uncontrolled Keywords: | Churn Management, Logistic regression, RSCV, SMOTE, RFE, Cross-Selling, Bundling. |
| Subjects: | Q Science > Q Science (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
| Divisions: | Fakultas Teknik > Program Studi Teknik Elektro (S1) |
| Depositing User: | MUHAMAD ILYAS HAIKAL |
| Date Deposited: | 19 Aug 2024 06:08 |
| Last Modified: | 28 Aug 2024 06:41 |
| URI: | http://repository.upnvj.ac.id/id/eprint/32375 |
Actions (login required)
 |
View Item |