Otkrivanje prijevara kreditnim karticama korištenjem varijacijskog autoenkodera

master's thesis

• Title (english): Credit card fraud detection using variational autoencoder
• Author: Damjan Grubelić
• Mentor: Zvonko Kostanjčar (mentor)
• Committee member: Zvonko Kostanjčar (predsjednik povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Committee member: Stjepan Begušić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-09-19, Croatia

Abstract

Otkrivanje prijevara jedan je od ključnih problema kod kartičnog plaćanja.
Naći skalabilan način za rješavanje tog problema nije jednostavan zadatak jer
broj legitimnih transakcija daleko nadmašuje broj prijevara. Osim toga, prevaranti
neprestano prilagođavaju pristupe pa detekcija mora biti robusna na promjene kroz
vrijeme.
Varijacijski autoenkoder može enkodirati distribucije iz kojih takvi podatci proizlaze
što omogućava robusnost na nesrazmjer u podatcima te generiranje novih, sličnih podataka
za smanjenje nesrazmjera. U ovom radu uspoređujem nekoliko inačica varijacijskog autoenkodera
te pristupa problemu detekcije prijevara pomoću njih.

Abstract (english)

Detecting fraud is one of the key issues in card payments. Finding a scalable way to address this problem is not an easy task because the number of legitimate transactions far exceeds the number of frauds. Moreover, fraudsters constantly adapt their approaches, so detection must be robust to changes over time.

A variational autoencoder can encode the distributions from which such data originate, enabling robustness to the imbalance in the data and the generation of new, similar data to reduce this imbalance. In this paper, I compare several variants of the variational autoencoder and their approaches to the problem of fraud detection.

• Keywords: VAE; generativni modeli; otkrivanje anomalija
• Keywords (english): VAE; generative models; anomaly detection
• Language: croatian
• URN:NBN: urn:nbn:hr:168:616797
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-11-07 05:37:45