Contents
How to measure class separability?
2.3.4 Class separability measures This can be done based on the absolute, Euclidean, Mahalanobis or nearest-neighbor distances, etc. The intraclass distances can be estimated by computing the average of absolute distances between the means of the class-conditional densities.
What is class separability?
The separability index measure estimates the average number of instances in a dataset that have a nearest neighbour with the same label. Since this is a fraction the index varies between 0-1 or 0-100%. Another separability measure, based on the class distance or margin is the Hypothesis margin (HM), introduced in [2].
How is class separability used in classification problems?
Class Separability: In classification problems, the reconstruction error is not the only criterion to measure the quality of the subspace. When label information is available, the discriminative ability is preferably taken into consideration. This information can be encoded in various ways.
Which is the distance metric for class separability?
Distance metric for measuring class separability: here, we have two class-conditional densities for two equiprobable classes. Notice that the classes can be partially separated based on the information about the class means. Classification errors are expected for features, x, that lie within the orange region. Fig. 2.6.
How to calculate separability of two dimensional data?
Scatter plots showing class separability of two-dimensional datasets. Notice that higher separability can be achieved by selecting those features that have large interclass means yet small within class variance.
How is clustering used to maximize class separability?
Criteria: Clustering can be considered as a technique to group samples so as to maximize class separability. Then, all of the criteria which were discussed in Chapter 10 may be used as clustering criteria. In this section only functions of scatter matrices are discussed due to the following reasons: