A simple and robust scoring technique for binary classification

Charles Gomes, Hisham Noçairi, Marie Thomas, Jean-François Collin, Gilbert Saporta

Abstract


A new simple scoring technique is developed in a binary supervised classification context when only a few observations areavailable. It consists in two steps: in the first one partial scores are obtained, one for each predictor, either categorical orcontinuous. Each partial score is a discrete variable with 7 values ranging from -3 to 3, based upon an empirical comparison ofthe distributions for each class. In a second step the partial scores are added and standardised into a global score, which allowsa decision rule.This simple technique is successfully compared with classical supervised techniques for a classical benchmark and has beenproved to be especially well fitted in an industrial problem.


Full Text: PDF DOI: 10.5430/air.v3n1p52

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

Artificial Intelligence Research

ISSN 1927-6974 (Print)   ISSN 1927-6982 (Online)

Copyright © Sciedu Press 
To make sure that you can receive messages from us, please add the 'Sciedu.ca' domain to your e-mail 'safe list'. If you do not receive e-mail in your 'inbox', check your 'bulk mail' or 'junk mail' folders.