Please use this identifier to cite or link to this item:
Title: The evidence-theoretic k-NN rule for rank-ordered data: Application to predict an individual’s source of loan
Authors: Supanika Leurcharusmee
Peerapat Jatukannyaprateep
Songsak Sriboonchitta
Thierry Denoeux
Keywords: Computer Science
Issue Date: 1-Jan-2014
Abstract: © Springer International Publishing Switzerland 2014. We adapted the nonparametric evidence-theoretic k-Nearest Neighbor (k-NN) rule,whichwas originally designed formultinomial choice data, to rank-ordered choice data.The contribution of thismodel is its ability to extract information from all the observed rankings to improve the prediction power for each individual’s primary choice. The evidence-theoretic k-NNrule for heterogeneous rank-ordered datamethod can be consistently applied to complete and partial rank-ordered choice data. This model was used to predict an individual’s source of loan given his or her characteristics and also identify individual characteristics that help the prediction. The results show that the prediction from the rank-ordered choice model outperforms that of the traditionalmultinomial choicemodelwith only one observed choice.
ISSN: 16113349
Appears in Collections:CMUL: Journal Articles

Files in This Item:
There are no files associated with this item.

Items in CMUIR are protected by copyright, with all rights reserved, unless otherwise indicated.