This paper describes the results of experimental investigation of two-feature evaluation criteria, i.e., inter-intra class distance ratio and information content measure. These two indirect statistical measures take into account higher order statistical redundancies among the feature being evaluated. The algorithms are first presented and then they are applied and compared to recognize handprinted alphanumeric characters. Both Highleyman's data and raw data obtained in the Signal Processing Laboratory at Case Western Reserve University, Cleveland, Ohio, were used for the study. It is believed that the criteria can be used for other applications and can especially be used where the statistical independency among features is not assumed.

Experimental Study of Information Measure and Inter-Intra Class Distance Ratios on Feature Selection and Orderings

Measurement of event-related potentials

Comments on "A modified figure of merit for feature selection in pattern recognition" by Paul, J.E., Jr., et al.

In this paper, measures of probabilistic dependence are derived from distance measures and are applied to feature evaluation in pattern recognition. The main properties of the measures are derived and are discussed in their application to feature-class dependency. Relations between the measures and error probability are derived. Experiments using feature subsets extracted from Munson's hand-printed data are performed to compare the feature-evaluating capabilities of the measures both relative to each other and relative to error probability.

Feature Evalution with Measures of Probabilistic Dependence

Cluster analysis of evoked potentials from behaving cats

The use of performance measures for selecting features in pattern recognition systems is reviewed. An approximation to the information content measure is derived. The approximation reduces the computation required to calculate the measure. The accuracy of the approximation depends directly on the nature of the patterns and their features. Computational requirements for the approximate measure are specified.

H. Ryan

Papers

The information content measure as a performance criterion for feature selection