Evaluasi Leakage-Aware dan Imbalance-Sensitive pada Algoritma Klasifikasi untuk Prediksi Keberhasilan Kampanye Bank Marketing
DOI:
https://doi.org/10.62712/juktisi.v5i1.1177Keywords:
bank marketing 1, classification 2,, class imbalance 3, arget leakage 4, threshold tuning 5,Abstract
Prediksi respons nasabah merupakan masalah penting dalam pemasaran perbankan berbasis data, terutama karena kampanye langsung harus menyeimbangkan efisiensi biaya, ketepatan penargetan, dan kemampuan mengenali calon nasabah yang benar-benar berpotensi merespons. Penelitian ini memperkuat evaluasi model prediksi kampanye Bank Marketing dengan dua prinsip metodologis, yaitu leakage-aware evaluation dan imbalance-sensitive evaluation. Atribut duration dikeluarkan dari model karena hanya diketahui setelah panggilan selesai sehingga berpotensi menimbulkan target leakage. Empat algoritma klasifikasi, yaitu Logistic Regression, K-Nearest Neighbor, Decision Tree, dan Random Forest, dievaluasi pada dataset bank.csv UCI yang berisi 4.521 observasi dengan distribusi kelas tidak seimbang, yaitu 4.000 kelas no dan 521 kelas yes. Eksperimen menggunakan train-validation-test stratified split, preprocessing berbasis standardisasi dan one-hot encoding, tuning hyperparameter melalui stratified cross-validation, serta evaluasi dengan accuracy, precision, recall, F1-score, F2-score, balanced accuracy, Matthews correlation coefficient, ROC-AUC, PR-AUC, dan confusion matrix. Selain evaluasi baseline pada threshold 0,50, penelitian ini juga menerapkan threshold tuning berbasis validasi dengan kriteria F2-score untuk meningkatkan sensitivitas terhadap kelas positif. Hasil menunjukkan bahwa Random Forest memiliki performa paling seimbang. Pada threshold 0,50, Random Forest memperoleh ROC-AUC 0,7576, PR-AUC 0,3743, MCC 0,2830, dan recall 0,4231. Setelah threshold dituning menjadi 0,39, recall Random Forest meningkat menjadi 0,7019 dengan F2-score 0,4980 dan balanced accuracy 0,6987. Temuan ini menunjukkan bahwa pemilihan model untuk kampanye pemasaran tidak cukup hanya berdasarkan accuracy, tetapi perlu mempertimbangkan trade-off antara recall, precision, false negative, dan tujuan operasional kampanye.
Downloads
References
S. Moro, P. Cortez, and P. Rita, "A data-driven approach to predict the success of bank telemarketing," Decision Support Systems, vol. 62, pp. 22-31, 2014, doi: 10.1016/j.dss.2014.03.001.
UCI Machine Learning Repository, "Bank Marketing Data Set," 2014. [Online]. Available: https://archive.ics.uci.edu/dataset/222/bank+marketing
G. James, D. Witten, T. Hastie, R. Tibshirani, and J. Taylor, An Introduction to Statistical Learning with Applications in R, 2nd ed. New York: Springer, 2021.
T. M. Cover and P. E. Hart, "Nearest neighbor pattern classification," IEEE Transactions on Information Theory, vol. 13, no. 1, pp. 21-27, 1967, doi: 10.1109/TIT.1967.1053964.
L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stone, Classification and Regression Trees. New York: Routledge, 1984.
L. Breiman, "Random Forests," Machine Learning, vol. 45, pp. 5-32, 2001, doi: 10.1023/A:1010933404324.
M. Sokolova and G. Lapalme, "A systematic analysis of performance measures for classification tasks," Information Processing and Management, vol. 45, no. 4, pp. 427-437, 2009, doi: 10.1016/j.ipm.2009.03.002.
F. Safarkhani and S. Moro, "Improving the Accuracy of Predicting Bank Depositor's Behavior Using a Decision Tree," Applied Sciences, vol. 11, no. 19, p. 9016, 2021, doi: 10.3390/app11199016.
C. Xie, J.-L. Zhang, Y. Zhu, B. Xiong, and G.-J. Wang, "How to improve the success of bank telemarketing? Prediction and interpretability analysis based on machine learning," Computers & Industrial Engineering, vol. 175, p. 108874, 2023, doi: 10.1016/j.cie.2022.108874.
W. Guo, Y. Yao, L. Liu, and T. Shen, "A novel ensemble approach for estimating the competency of bank telemarketing," Scientific Reports, vol. 13, p. 20819, 2023, doi: 10.1038/s41598-023-47177-7.
F. Nasir, A. A. Ahmed, M. S. Kiraz, I. Yevseyeva, and M. Saif, "Data-Driven Decision-Making for Bank Target Marketing Using Supervised Learning Classifiers on Imbalanced Big Data," Computers, Materials & Continua, vol. 81, no. 1, pp. 1703-1728, 2024, doi: 10.32604/cmc.2024.055192.
M. Peter, H. Mofi, S. Likoko, J. Sabas, R. Mbura, and N. Mduma, "Predicting customer subscription in bank telemarketing campaigns using ensemble learning models," Machine Learning with Applications, vol. 19, p. 100618, 2025, doi: 10.1016/j.mlwa.2025.100618.
N. V. Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer, "SMOTE: Synthetic Minority Over-sampling Technique," Journal of Artificial Intelligence Research, vol. 16, pp. 321-357, 2002, doi: 10.1613/jair.953.
G. Lemaitre, F. Nogueira, and C. K. Aridas, "Imbalanced-learn: A Python Toolbox to Tackle the Curse of Imbalanced Datasets in Machine Learning," Journal of Machine Learning Research, vol. 18, no. 17, pp. 1-5, 2017.
F. Pedregosa et al., "Scikit-learn: Machine Learning in Python," Journal of Machine Learning Research, vol. 12, pp. 2825-2830, 2011.
T. Fawcett, "An introduction to ROC analysis," Pattern Recognition Letters, vol. 27, no. 8, pp. 861-874, 2006, doi: 10.1016/j.patrec.2005.10.010.
J. Davis and M. Goadrich, "The relationship between Precision-Recall and ROC curves," in Proceedings of the 23rd International Conference on Machine Learning, 2006, pp. 233-240, doi: 10.1145/1143844.1143874.
D. Chicco and G. Jurman, "The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation," BMC Genomics, vol. 21, no. 1, p. 6, 2020, doi: 10.1186/s12864-019-6413-7.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Purwatiningtyas, Retnowati

This work is licensed under a Creative Commons Attribution 4.0 International License.















