Naive Bayes classifier for positive unlabeled learning with uncertainty

He, Jiazhen, Zhang, Yang, Li, Xue and Wang, Yong (2010). Naive Bayes classifier for positive unlabeled learning with uncertainty. In: Proceedings of the Tenth SIAM International Conference on Data Mining. The Tenth SIAM International Conference on Data Mining (SDM10), Columbus, Ohio, United States, (361-372). 29 April - 1 May 2010.

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
Preface_SDM10_Proceedings.mht ERA evidence - not publicly available Click to show the corresponding preview/stream 157.84KB 1879
UQ233589_fulltext.pdf ERA evidence - not publicly available application/pdf 782.78KB 82
Author He, Jiazhen
Zhang, Yang
Li, Xue
Wang, Yong
Title of paper Naive Bayes classifier for positive unlabeled learning with uncertainty
Conference name The Tenth SIAM International Conference on Data Mining (SDM10)
Conference location Columbus, Ohio, United States
Conference dates 29 April - 1 May 2010
Proceedings title Proceedings of the Tenth SIAM International Conference on Data Mining
Journal name Proceedings of the 10th SIAM International Conference on Data Mining, SDM 2010
Place of Publication Philadelphia, PA, United States
Publisher Society for Industrial and Applied Mathematics (SIAM)
Publication Year 2010
Sub-type Fully published paper
Open Access Status File (Author Post-print)
Start page 361
End page 372
Total pages 12
Language eng
Formatted Abstract/Summary
Existing algorithms for positive unlabeled learning (PU
learning) only work with certain data. However, data un-
certainty is prevalent in many real-world applications such
as sensor network, market analysis and medical diagnosis.
In this paper, based on positive naive Bayes (PNB), which
is a PU learning algorithm for certain data, we propose an
algorithm to handle uncertain data . However, it requires
the prior probability of positive class and in real-life ap-
plications it is generally difficult for the users to provide
this parameter, which is a drawback inherited from tradi-
tional PNB algorithm. We improve it by selecting the value
of the prior probability of positive class automatically that
can make the obtained classifier achieved optimal perfor-
mance on the validation set. The conducted experiments
show that the proposed algorithm yields good performance
without user-specified the prior probability of positive class
and has satisfactory performance even on highly uncertain
data.
Q-Index Code E1
Q-Index Status Provisional Code
Institutional Status UQ

 
Versions
Version Filter Type
Citation counts: Scopus Citation Count Cited 0 times in Scopus Article
Google Scholar Search Google Scholar
Created: Wed, 09 Mar 2011, 10:28:48 EST by Dr Xue Li on behalf of School of Information Technol and Elec Engineering