PREDIKSI CUSTOMER CHURN PADA BANK MENGGUNAKAN PERBANDINGAN OPTIMASI ALGORITMA GRID SEARCH DENGAN DECISION TREE, RANDOM FOREST, XGBOOST

Sari, Sabrina Laila (2025) PREDIKSI CUSTOMER CHURN PADA BANK MENGGUNAKAN PERBANDINGAN OPTIMASI ALGORITMA GRID SEARCH DENGAN DECISION TREE, RANDOM FOREST, XGBOOST. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) Buku Skripsi Baru (1).pdf Download (1MB)
	Text (Bab 1) Bab 1.pdf Download (830kB)
	Text (Bab 2) Bab 2.pdf Restricted to Repository staff only until 12 December 2028. Download (2MB)
	Text (Bab 3) Bab 3.pdf Restricted to Repository staff only until 12 December 2028. Download (3MB)
	Text (Bab 4) Bab 4.pdf Restricted to Repository staff only until 12 December 2028. Download (10MB)
	Text (Bab 5) Bab 5.pdf Download (416kB)
	Text (Daftar Pustaka) Daftar Pustaka.pdf Download (714kB)
	Text (Lampiran) Lampiran.pdf Restricted to Repository staff only until 12 December 2028. Download (127kB)

Abstract

This study applies Grid Search hyperparameter optimization to improve machine learning models for predicting customer churn in the banking sector. Using a publicly available dataset of 10,000 customers with 14 attributes, the workflow included data cleaning and transformation (encoding, normalization), train–test splitting (70:30 and 80:20), model training, and exhaustive hyperparameter tuning. Three tree-based classifiers were compared: Decision Tree, Random Forest, and XGBoost. Model performance was evaluated with accuracy, precision, recall, F1-score, ROC AUC, and PR AUC. Results show that after Grid Search tuning all three models achieved approximately 85% accuracy in the best scenarios, while ensemble methods outperformed the single tree in distinguishing churn cases. Random Forest and XGBoost produced higher ROC AUC and PR AUC values than Decision Tree, and XGBoost delivered the best balance between precision and recall for the minority churn class after tuning. Feature analysis identified age, active membership status, number of products, and account balance as the most influential predictors. Grid Search significantly improved model effectiveness but increased computational cost; future work should explore more efficient optimization methods (e.g., randomized search, Bayesian optimization) and additional feature engineering to further enhance churn detection.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Rahmat, Basuki	NIP196907232021211002	basukirahmat.if@upnjatim.ac.id
Thesis advisor	Kartini, Kartini	NIP196111101991032001	kartini.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA75 Electronic computers. Computer science

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Unnamed user with email 20081010224@student.upnjatim.ac.id

Date Deposited:

22 Jan 2026 05:59

Last Modified:

22 Jan 2026 07:00

URI:

https://repository.upnjatim.ac.id/id/eprint/48788