Skip to main navigation Skip to search Skip to main content

Naïve Bayes vs. support vector machine: Resilience to missing data

  • Hongbo Shi*
  • , Yaqin Liu
  • *Corresponding author for this work
  • Shanxi University of Finance and Economics

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The naïve Bayes and support vector machine are the typical generative and discriminative classification models respectively, which are two popular classification approaches. Few studies have been done comparing their resilience to missing data. This paper provides an experimental comparison of the naïve Bayes and support vector machine regarding the resilience to missing data on 24 UCI data sets. The experimental results show that when the missing rate is very small (e.g. 1%), the resilience of the naïve Bayes classifiers to missing data are approximately similar to that of support vector machine classifiers. With the increase of the missing rate, however, the resilience of the naïve Bayes classifiers to missing data are slowly decreased and that of support vector machine classifiers to missing data are rapidly decreased. This demonstrates that the naïve Bayes classifiers have better resilience to missing data than support vector machine classifiers.

Original languageEnglish
Title of host publicationArtificial Intelligence and Computational Intelligence - Third International Conference, AICI 2011, Proceedings
Pages680-687
Number of pages8
EditionPART 2
DOIs
StatePublished - 2011
Externally publishedYes
Event3rd International Conference on Artificial Intelligence and Computational Intelligence, AICI 2011 - Taiyuan, China
Duration: 24 Sep 201125 Sep 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume7003 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd International Conference on Artificial Intelligence and Computational Intelligence, AICI 2011
Country/TerritoryChina
CityTaiyuan
Period24/09/1125/09/11

Keywords

  • SVM
  • missing data
  • resilience
  • the naïve Bayes

Fingerprint

Dive into the research topics of 'Naïve Bayes vs. support vector machine: Resilience to missing data'. Together they form a unique fingerprint.

Cite this