A study of the robustness of KNN classifiers trained using soft labels

Neamat El Gayar, Friedhelm Schwenker, Günther Palm

Research output: Chapter in Book/Report/Conference proceedingConference contribution

29 Citations (Scopus)

Abstract

Supervised learning models most commonly use crisp labels for classifier training. Crisp labels fail to capture the data characteristics when overlapping classes exist. In this work we attempt to compare between learning using soft and hard labels to train K-nearest neighbor classifiers. We propose a new technique to generate soft labels based on fuzzy-clustering of the data and fuzzy relabelling of cluster prototypes. Experiments were conducted on five data sets to compare between classifiers that learn using different types of soft labels and classifiers that learn with crisp labels. Results reveal that learning with soft labels is more robust against label errors opposed to learning with crisp labels. The proposed technique to find soft labels from the data, was also found to lead to a more robust training in most data sets investigated.

Original languageEnglish
Title of host publicationArtificial Neural Networks in Pattern Recognition. ANNPR 2006
PublisherSpringer
Pages67-80
Number of pages14
ISBN (Electronic)9783540379522
ISBN (Print)9783540379515
DOIs
Publication statusPublished - 2006
Event2nd IAPR Workshop on Artificial Neural Networks in Pattern Recognition 2006 - Ulm, Germany
Duration: 31 Aug 20062 Sep 2006

Publication series

NameLecture Notes in Computer Science
Volume4087
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd IAPR Workshop on Artificial Neural Networks in Pattern Recognition 2006
Abbreviated titleANNPR 2006
CountryGermany
CityUlm
Period31/08/062/09/06

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'A study of the robustness of KNN classifiers trained using soft labels'. Together they form a unique fingerprint.

  • Cite this

    El Gayar, N., Schwenker, F., & Palm, G. (2006). A study of the robustness of KNN classifiers trained using soft labels. In Artificial Neural Networks in Pattern Recognition. ANNPR 2006 (pp. 67-80). (Lecture Notes in Computer Science; Vol. 4087). Springer. https://doi.org/10.1007/11829898_7