Screening Covariates in Presence of Unbalanced Binary Dependent Variable

Francesco Giordano (), Marcella Niglio () and Marialuisa Restaino ()
Additional contact information
Francesco Giordano: University of Salerno
Marcella Niglio: University of Salerno

A chapter in Mathematical and Statistical Methods for Actuarial Sciences and Finance, 2021, pp 257-263 from Springer

Abstract: Abstract In this contribution we propose a new method to identify the most relevant covariates in a large dataset that (i) is applicable in presence of regression models where the binary dependent variable is characterized by a very small number of ones than zeros, (ii) is not strongly influenced by the correlation between covariates, (iii) is easily applied when the number of predictors increases up to infinity and/or it is greater than the sample size. The proposed procedure extends the idea of Sure Independence Screening for the linear regression model to the Generalized Extreme Value regression framework. This technique allows to define a set of relevant covariates that survive after applying the screening procedure and that, with a probability tending to one, includes the true relevant covariates. We validate the proposed procedure by a simulation study and an empirical analysis devoted to the prediction of firms failure.

Keywords: GEV; Screening; Variable selection (search for similar items in EconPapers)
Date: 2021
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-3-030-78965-7_38

Ordering information: This item can be ordered from
http://www.springer.com/9783030789657

DOI: 10.1007/978-3-030-78965-7_38

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().