Modeli dubokog učenja za procjenu rizika izravnog kreditiranja

master's thesis

• Title (english): Deep Learning Models for Risk Assessment in Peer-to-Peer Lending
• Author: Ivan Ucović
• Mentor: Zvonko Kostanjčar (mentor)
• Committee member: Zvonko Kostanjčar (predsjednik povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Committee member: Juraj Petrović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Modeli dubokog učenja za procjenu rizika izravnog kreditiranja
U ovom radu razvijeni su modeli strojnog i dubokog učenja za predikciju kreditnog rizika izravnog ili P2P kreditiranja. Naime izravno kreditiranje nije tako novi trend u financijskoj industriji, međutim jedan je od najbrže rastućih u istoj. Ovakvo kreditiranje sadrži rizike za kreditore, koji posuđuju vlastiti novac, pa se pojavljuje sve veća potreba za kvalitetnim modelima procjene kreditnog rizika. U ovom radu razvijeni su modeli koji koristeći podatke koji su dostupni kreditorima prije odobravanje kredita, diskriminiraju dobre od loših kredita. Modeli su evaluirani nizom metrika koje su standard u industriji.

Abstract (english)

Deep models for credit risk assessment in peer to peer financing.
In this paper, three models for credit risk assessment were created using machine and deep learning techniques. Peer to peer financing is not that new, but it is the fastest growing industry in fintech. This type of crediting involves huge risks for investors since they are borowing their own money. This poses a new need for quality credit risk assessment models. Models in this paper were created using dana available to investors prior to credit acceptance and were trained to differerntiate between good and bad credits. At the end these models were evaluated by roc-auc metric, as a standard metric in this industry as well as some other know metrics.

• Keywords: Duboko učenje; strojno učenje; neuronske mreže; kreditni rizik; izravno financiranje
• Keywords (english): Deep Learning; neural networks; machine learning; credit risk; peer to peer finance
• Language: croatian
• URN:NBN: urn:nbn:hr:168:994092
• 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: 2023-01-13 14:06:51