Sensitivity and specificity for mining data with increased incompleteness

  • Authors:
  • Jerzy W. Grzymala-Busse;Shantan R. Marepally

  • Affiliations:
  • Department of Electrical Engineering and Computer Science, University of Kansas, Lawrence, KS and Institute of Computer Science, Polish Academy of Sciences, Warsaw, Poland;Department of Electrical Engineering and Computer Science, University of Kansas, Lawrence, KS

  • Venue:
  • ICAISC'10 Proceedings of the 10th international conference on Artificial intelligence and soft computing: Part I
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents results of experiments on data sets that were subjected to increasing incompleteness by random replacement of attribute values by symbols of missing attribute values. During these experiments the total error rate and error rates for all concepts, results of repeated 30 times ten-fold cross validation, were recorded. We observed that for some data sets increased incompleteness might result in a significant improvement for the total error rate and sensitivity (with the significance level of 5%, two-tailed test). These results may be applied for improving data mining techniques, especially for domains in which sensitivity is important, e.g., in medical area.