Probabilistic Distance Measure Research Articles

A Probabilistic Distance Measure for Hidden Markov Models

We propose a probabilistic distance measure for measuring the dissimilarity between pairs of hidden Markov models with arbitrary observation densities. The measure is based on the Kullback-Leibler number and is consistent with the reestimation technique for hidden Markov models. Numerical examples that demonstrate the utility of the proposed distance measure are given for hidden Markov models with discrete densities. We also discuss the effects of various parameter deviations in the Markov models on the resulting distance, and study the relationships among parameter estimates (obtained from reestimation), initial guesses of parameter values, and observation duration through the use of the measure.

Efficient feature-subset selection with probabilistic distance criteria

This paper considers the problem of finding the best feature subset by exhaustive search, using probabilistic distance measures as criteria. Recursive expressions are derived for efficiently computing the commonly used probabilistic distance measures as a change in the criteria both when a feature is added to and when a feature is deleted from the current feature subset. A combinatorial algorithm is presented for generating all possible r-feature combinations from a given set of S features in ( S r ) steps with a change of a single feature at each step. These expressions can also be used for both forward and backward sequential feature selection.

Comment on "The relationship between the equivocation and other probabilistic distance measures used for feature selection"

Comment on "The relationship between the equivocation and other probabilistic distance measures used for feature selection"

On the application of probabilistic distance measures for the extraction of features from imperfectly labeled patterns

A commonly used approach for feature selection is to select those features that extremize certain probabilistic distance measures. In most of the procedures it is assumed that the labels of the patterns are perfect. There are many practical situations in which the labels of the patterns are imperfect. This paper examines the applicability of the extremization of the Bhattacharyya distance, the divergence, equivocation, Kalmogrov variational distance, and Matusita distance as criteria for selecting the effective features from imperfectly identified patterns.

On the relationship between the equivocation and other probabilistic distance measures used for feature selection

Some inequalities are derived between the equivocation, the Bhattacharyya coefficient, the divergence, the Kalmogrov variational distance, and the Matusita distance.