Mining Mass Spectrometry Database Search Results--A Rough Set Approach

Authors:
Jianwen Fang;Jerzy W. Grzymala-Busse
Affiliations:
Bioinformatics Core Facility, and Information and Telecommunication Technology Center, University of Kansas, Lawrence, KS 66045, USA;Department of Electrical Engineering and Computer Science, University of Kansas, Lawrence, KS 66045, USA, and Institute of Computer Science Polish Academy of Sciences, 01-237 Warsaw, Poland
Venue:
RSEISP '07 Proceedings of the international conference on Rough Sets and Intelligent Systems Paradigms
Year:
2007

Citing 6
Cited 0

Induction: processes of inference, learning, and discovery

Induction: processes of inference, learning, and discovery
Classifier systems and genetic algorithms

Machine learning: paradigms and methods
A new version of the rule induction system LERS

Fundamenta Informaticae
Rough Sets: Theoretical Aspects of Reasoning about Data

Rough Sets: Theoretical Aspects of Reasoning about Data
Postprocessing of Rule Sets Induced from a Melanoma Data Set

COMPSAC '02 Proceedings of the 26th International Computer Software and Applications Conference on Prolonging Software Life: Development and Redevelopment
TANDEM: matching proteins with tandem mass spectra

Bioinformatics

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper reports results of experiments on mass spectrometry database search results produced by Keller et al. This data set describes human proteins. Data mining was conducted using the LERS system. First, the data set was discretized by a cluster analysis algorithm based on agglomerative approach. Then the basic rule set was induced by the LEM2 algorithm. Finally, the rule set was refined using changing rule strength methodology and truncation of the rule set. Our results reach the level of sensitivity and specificity of competing methods. However, our results are explainable since they are in a form of rules and, additionally, we can interpret the role of important features.