Deteksi Penipuan Pada Transaksi Kartu Kredit Menggunakan Metode Stochastic Gradient Descent Dengan Momentum

Mandalla, Achmad Zaki (2023) Deteksi Penipuan Pada Transaksi Kartu Kredit Menggunakan Metode Stochastic Gradient Descent Dengan Momentum. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06111540000095-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (3MB) | Request a copy

Abstract

Penipuan kartu kredit memberikan kerugian yang sangat besar bagi pemilik kartu, pemberi kredit, dan pihak terkait. Permasalahan ini terus meningkat setiap tahunnya dan telah menjadi permasalahan yang serius. Dalam penelitian ini, dilakukan proses data mining untuk mendapatkan pola atau informasi yang berguna berupa klasifikasi penipuan transaksi kartu kredit. Secara khusus, metode yang digunakan adalah stochastic gradient descent dengan momentum (SGDM). Selanjutnya dilakukan perbandingan hasil klasifikasi menggunakan metode stochastic gradient descent dengan momentum (SGDM), gradient descent (GD), dan stochastic gradient descent (SGD). Pengujian tersebut menunjukkan bahwa metode SGDM memiliki hasil klasifikasi yang lebih baik dari metode GD dan SGD dengan nilai akurasi 99,85%, f1-score 90,15%, precision 91%, dan recall 90%.
=====================================================================================================================================
Credit card fraud causes huge losses to cardholders, lenders, and related parties. This problem continues to increase every year and has become a serious problem. In this research, a data mining process is carried out to obtain useful patterns or information in the form of credit card transaction fraud classification. In particular, the method used is stochastic gradient descent with momentum (SGDM). Furthermore, a comparison of classification results using stochastic gradient descent with momentum method (SGDM), gradient descent method (GD), and stochastic gradient descent method (SGD). The test shows that SGDM method has better classification results than GD and SGD methods with an accuracy value of 99.85%, f1-score 90.15%, precision 91%, and recall 90%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	data mining, klasifikasi, penipuan kartu kredit, stochastic gradient descent dengan momentum, data mining, classification, credit card fraud, stochastic gradient descent with momentum.
Subjects:	Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Achmad Zaki Mandalla
Date Deposited:	16 Feb 2023 06:13
Last Modified:	16 Feb 2023 06:13
URI:	http://repository.its.ac.id/id/eprint/97350

Actions (login required)

View Item