I discuss diagnostic methods for discriminant analysis. The equivalence with linear regression is noted and regression diagnostics are considered. The leverage is a function of the linear discriminant function and the Mahalanobis distance of the observation from the group mean. The distribution of this distance is approximately chi-square with degrees of freedom equal to the number of variables...

Representing motions as linear sums of principal components has become a widely accepted animation technique. While powerful, the simplest version of this approach is not particularly well suited to modeling the specific style of an individual whose motion had not yet been recorded when building the database: It would take an expert to adjust the PCA weights to obtain a motion style that is ind...

The accuracy of the k-nearest neighbor algorithm depends on the distance function used to measure similarity between instances. Methods have been proposed in the literature to learn a good distance function from a labelled training set. One such method is the large margin nearest neighbor classifier that learns a global Mahalanobis distance. We propose a mixture of such classifiers where a gati...

Atypical observations hidden in the data may play quite an disastrous role in a tted regression, especially when commonly used outlier detection techniques like computing leverages, Mahalanobis distances, ordinary and studentized residuals, DFFits, cross-validations { do not detect them. However (multivariate) outliers can be detected quite easily by graphical techniques , e.g. scatterplot matr...