Measuring and Mitigating Bias in Bank Customers Data with XGBoost, LightGBM, and Random Forest Algorithm

Wardani, Berliana Shafa and Sa’adah, Siti and Nurjanah, Dade (2023) Measuring and Mitigating Bias in Bank Customers Data with XGBoost, LightGBM, and Random Forest Algorithm. Jurnal Ilmiah Teknik Elektro Komputer dan Informatika (JITEKI), 9 (1). pp. 142-155.

[thumbnail of 13-Measuring and Mitigating Bias in Bank Customers Data with XGBoost, LightGBM, and Random Forest Algorithm.pdf] Text
13-Measuring and Mitigating Bias in Bank Customers Data with XGBoost, LightGBM, and Random Forest Algorithm.pdf

Download (781kB)

Abstract

To retain its customers, Portuguese banking institutions carry out direct marketing in the form of telephone calls to conduct marketing so that customers subscribe to the bank's term deposits. This research was conducted with bank customer data from a Portuguese banking institution that implemented agent acquisition. The problem is that the large amount of bank customer data can cause data diversity which allows the results of agent acquisition to be unfair so that the features in the data must really be considered in the acquisition process. For example, gender inequality in data can cause decision results to be skewed to one group so that other groups are disadvantaged. Thus, a bias detection and mitigation algorithm is needed to achieve fairness so as to produce better predictive results. AI fairness 360 (AIF 360) is a toolkit that provides bias detection and mitigation algorithms. The bias mitigation algorithm in AIF 360 is divided into three processes, namely reweighing and learning fair representation at the pre-processing stage, debunking and debasing hostility at the in-processing stage, and classification of equalized odds and reject options at the post-processing stage. The output of this study is a comparison of the calculation of bias detection with different impacts (DI) and statistical parity differences (SPD) before and after mitigation. The adversarial debiasing algorithm performs better than others with DI 0.943, SPD -0.004, and also increases the AUC score by 0.015%. Doing this research can help predict customer deposits in Portuguese banking institutions more fairly.

Item Type: Artikel Umum
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisi / Prodi: Faculty of Industrial Technology (Fakultas Teknologi Industri) > S1-Electrical Engineering (S1-Teknik Elektro)
Depositing User: M.Eng. Alfian Ma'arif
Date Deposited: 24 Mar 2023 03:41
Last Modified: 24 Mar 2023 03:41
URI: http://eprints.uad.ac.id/id/eprint/41476

Actions (login required)

View Item View Item