OpenstarTs >
EUT-Libri >
Collane >
Working Paper Series - Dipartimento di scienze economiche, aziendali, matematiche e statistiche "Bruno de Finetti" >
Working Papers Series 2010, 2 >

Please use this identifier to cite or link to this item:

Title: Training and assessing classification rules with unbalanced data
Authors: Menardi, Giovanna
Torelli, Nicola
Keywords: accuracy
binary classification
kernel density estimation
unbalanced learning
Issue Date: 2010
Publisher: EUT Edizioni Università di Trieste
Citation: Giovanna Menardi, Nicola Torelli, "Training and assessing classification rules with unbalanced data", Working Paper Series, N. 2, 2010.
Series/Report no.: Working paper series - Dipartimento di scienze economiche, aziendali, matematiche e statistiche "Bruno de Finetti"
2 (2010)
Abstract: The problem of modeling binary responses by using cross-sectional data has been addressed with a number of satisfying solutions that draw on both parametric and nonparametric methods. However, there exist many real situations where one of the two responses (usually the most interesting for the analysis) is rare. It has been largely reported that this class imbalance heavily compromises the process of learning, because the model tends to focus on the prevalent class and to ignore the rare events. However, not only the estimation of the classification model is affected by a skewed distribution of the classes, but also the evaluation of its accuracy is jeopardized, because the scarcity of data leads to poor estimates of the model’s accuracy. In this work, the effects of class imbalance on model training and model assessing are discussed. Moreover, a unified and systematic framework for dealing with both the problems is proposed, based on a smoothed bootstrap re-sampling technique.
ISBN: 978-88-8303-321-6
Appears in Collections:Working Papers Series 2010, 2

Files in This Item:

File Description SizeFormat
Menardi Torelli DEAMS WPS2.pdf503.89 kBAdobe PDFView/Open
View Statistics

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