DETEKSI ANOMALI MENGGUNAKAN ENSEMBLE LEARNING DAN RANDOM OVERSAMPLING PADA PENIPUAN TRANSAKSI KEUANGAN

Saputra, Dewa Raka Krisna (2024) DETEKSI ANOMALI MENGGUNAKAN ENSEMBLE LEARNING DAN RANDOM OVERSAMPLING PADA PENIPUAN TRANSAKSI KEUANGAN. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 20081010205.-cover.pdf Download (2MB)
	Text (BAB 1) 20081010205.-bab1.pdf Download (850kB)
	Text (BAB 2) 20081010205.-bab2.pdf Restricted to Repository staff only until 19 July 2026. Download (4MB)
	Text (BAB 3) 20081010205.-bab3.pdf Restricted to Repository staff only until 19 July 2026. Download (8MB)
	Text (BAB 4) 20081010205.-bab4.pdf Restricted to Repository staff only until 19 July 2026. Download (10MB)
	Text (BAB 5) 20081010205.-bab5.pdf Download (312kB)
	Text (DAFTAR PUSTAKA) 20081010205_daftarpustaka.pdf Download (191kB)

Abstract

In the digital era, financial transactions are increasingly turning to non-cash methods, because they are convenient and efficient. However, increased use of credit cards and online transactions also increases the risk of financial crime. This research examines ensemble learning and random oversampling methods in detecting anomalies in financial transactions, especially credit card fraud. The classification algorithms used include Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), and Naive Bayes (NB), with ensemble learning approaches such as Bagging, Boosting, and Stacking. The research results show that the ensemble learning method significantly improves fraud detection performance compared to the base model. In particular, stacking techniques show significant AUC improvements, with some algorithms achieving perfect AUC (1.00). Random Forest (RF) with the ensemble learning method shows very consistent and optimal performance in detecting fraud anomalies. This research confirms that ensemble learning methods, especially stacking, are effective in distinguishing between legitimate and suspicious transactions, making them reliable for financial fraud detection.

Item Type:

Thesis (Undergraduate)

Contributors: