Discovering rules-based similarity in microarray data

  • Authors:

  • Andrzej Janusz

  • Affiliations:

  • Faculty of Mathematics, Informatics, and Mechanics, The University of Warsaw, Warszawa, Poland

  • Venue:
  • IPMU'10 Proceedings of the Computational intelligence for knowledge-based systems design, and 13th international conference on Information processing and management of uncertainty
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a research on discovering a similarity relation in multidimensional bioinformatic data. In particular, utilization of a Rules-based Similarity model to define a similarity in microarray datasets is discussed. The Rules-based Similarity model is a rough set extension to the feature contrast model proposed by Amos Tversky. Its main aim is to achieve high accuracy in a case-based classification task and at the same time to simulate the human way of perceiving similar objects. The similarity relation derived from the Rules-based Similarity model is suitable for genes expression profiling as the rules naturally indicate the groups of genes whose activation or inactivation is relevant in the considered context. Experiments conducted on several microarray datasets show that this model of similarity is able to capture higher-level dependencies in data and it may be successfully used in cases when the standard distance-based approach turns out to be ineffective.