dc.contributor.author | Andrade Mancisidor, Rogelio | |
dc.contributor.author | Kampffmeyer, Michael | |
dc.contributor.author | Aas, Kjersti | |
dc.contributor.author | Jenssen, Robert | |
dc.date.accessioned | 2022-08-11T11:55:13Z | |
dc.date.available | 2022-08-11T11:55:13Z | |
dc.date.issued | 2022-03-17 | |
dc.description.abstract | Banks collect data x<sub>1</sub> in loan applications to decide whether to grant credit and accepted applications
generate new data x<sub>2</sub> throughout the loan period. Hence, banks have two measurement-modalities,
which provide a complete picture about customers. If we can generate x<sub>2</sub> conditioned on x<sub>1</sub> keeping the
relationship between these two modalities, credit and behavior scoring may be enabled simultaneously
(at the time x<sub>1</sub> is obtained) to support cross-selling, launching of new products or marketing
campaigns. Therefore, we develop a novel conditional bi-modal discriminative (CBMD) model for credit
scoring, which is able to generate x<sub>2</sub> based on x<sub>1</sub> and can classify the outcome of loans in an unified
framework. The idea behind CBMD is to learn joint (among modalities) latent representations that
are useful to generate x<sub>2</sub> using the available data x1 during the application process. The classifier
model introduced in CBMD encourages the generative process to generate x<sub>2</sub> accurately. Further, CBMD
optimizes a novel objective function introduced in this research, which maximizes mutual information
between the latent representation z and the modality x<sub>2</sub> to improve the generative process in the
model. We benchmark the generative process of our proposed model and CBMD outperforms other
multi-learning models. Similarly, the classification performance of CBMD is tested under different
scenarios and it achieves higher or on a par model performance compared to the state-of-the-art
in multi-modal learning models. | en_US |
dc.identifier.citation | Andrade Mancisidor, Kampffmeyer, Aas, Jenssen. Generating customer's credit behavior with deep generative models. Knowledge-Based Systems. 2022;245:1-13 | en_US |
dc.identifier.cristinID | FRIDAID 2019914 | |
dc.identifier.doi | 10.1016/j.knosys.2022.108568 | |
dc.identifier.issn | 0950-7051 | |
dc.identifier.issn | 1872-7409 | |
dc.identifier.uri | https://hdl.handle.net/10037/26148 | |
dc.language.iso | eng | en_US |
dc.publisher | Elsevier | en_US |
dc.relation.journal | Knowledge-Based Systems | |
dc.rights.accessRights | openAccess | en_US |
dc.rights.holder | Copyright 2022 The Author(s) | en_US |
dc.title | Generating customer's credit behavior with deep generative models | en_US |
dc.type.version | publishedVersion | en_US |
dc.type | Journal article | en_US |
dc.type | Tidsskriftartikkel | en_US |
dc.type | Peer reviewed | en_US |