Pregled bibliografske jedinice broj: 1258943
Default Prediction Using Consumption Data
Default Prediction Using Consumption Data // 2022 International Conference on Computing, Networking, Telecommunications & Engineering Sciences Applications (CoNTESA)
Skopje, Sjeverna Makedonija: Institute of Electrical and Electronics Engineers (IEEE), 2022. str. 11-16 doi:10.1109/contesa57046.2022.10011333 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 1258943 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Default Prediction Using Consumption Data
Autori
Mercep, Andro ; Kovacevic, Tomislav ; Bauman, Tessa ; Kostanjcar, Zvonko
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
2022 International Conference on Computing, Networking, Telecommunications & Engineering Sciences Applications (CoNTESA)
/ - : Institute of Electrical and Electronics Engineers (IEEE), 2022, 11-16
Skup
3rd International Conference on Computing, Networking, Telecommunications & Engineering Sciences Applications 2022 (CoNTESA '22)
Mjesto i datum
Skopje, Sjeverna Makedonija, 15.12.2022. - 16.12.2022
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
consumption data ; credit risk assessment ; machine learning ; probability of default
Sažetak
Current practice in the development of credit risk assessment models relies on a large amount of client data, such as account balance information, socio-demographic data, tenure features, or credit history of the client. With the increasing number of different data sources, it is possible to employ nonconventional information to assess clients’ credit risk. Consumption data are an example of such a source, as they provide an overview of clients’ spending habits. However, considering this information at a level of individual companies where clients’ purchases took place inherently leads to sparse data, as customers typically shop at only a small subset of the available retailers. We demonstrate that commonly used models suffer from poor performance caused by the sparsity of the consumption data, and applying dimensionality reduction techniques such as PCA does not offer a meaningful improvement. In this paper, we propose two word2vec based models that leverage the contextual information of clients’ spending habits. Both of the developed models achieved significantly higher Gini scores, outperforming the baseline models by as much as 72%.
Izvorni jezik
Engleski
Znanstvena područja
Računarstvo
POVEZANOST RADA
Projekti:
HRZZ-IP-2019-04-5241 - Algoritmi dubokog podržanog učenja za upravljanje rizicima (DREAM) (Kostanjčar, Zvonko, HRZZ ) ( CroRIS)
KK.01.1.1.01.0009 - Napredne metode i tehnologije u znanosti o podatcima i kooperativnim sustavima (EK )
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb
