Effects of Sample Size in Classifier Design

  • Authors:
  • K. Fukunaga;R. R. Hayes

  • Affiliations:
  • Purdue Univ., West Lafayette, IN;Purdue Univ., West Lafayette, IN

  • Venue:
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Year:
  • 1989

Quantified Score

Hi-index 0.16

Visualization

Abstract

The effect of finite sample-size on parameter estimates and their subsequent use in a family of functions are discussed. General and parameter-specific expressions for the expected bias and variance of the functions are derived. These expressions are then applied to the Bhattacharyya distance and the analysis of the linear and quadratic classifiers, providing insight into the relationship between the number of features and the number of training samples. Because of the functional form of the expressions, an empirical approach is presented to enable asymptotic performance to be accurately estimated using a very small number of samples. Results were experimentally verified using artificial data in controlled cases and using real, high-dimensional data.