In high dimension, low sample size (HDLSS) settings, classifiers based on Euclidean distances like the nearest neighbor classifier and the average distance classifier perform quite poorly if differences between locations of the underlying populations get masked by scale differences. To rectify this problem, several modifications of these classifiers have been proposed in the literature. However, existing methods are confined to location and scale differences only, and they often fail to discriminate among populations differing outside of the first two moments. In this article, we propose some simple transformations of these classifiers resulting in improved performance even when the underlying populations have the same location and scale. We further propose a generalization of these classifiers based on the idea of grouping of variables. High-dimensional behavior of the proposed classifiers is studied theoretically. Numerical experiments with a variety of simulated examples as well as an extensive analysis of benchmark data sets from three different databases exhibit advantages of the proposed methods.

Popular clustering algorithms based on usual distance functions (e.g., the Euclidean distance) often suffer in high dimension, low sample size (HDLSS) situations, where concentration of pairwise distances and violation of neighborhood structure have adverse effects on their performance. In this article, we use a new data-driven dissimilarity measure, called MADD, which takes care of these problems. MADD uses the distance concentration phenomenon to its advantage, and as a result, clustering algorithms based on MADD usually perform well for high dimensional data. We establish it using theoretical as well as numerical studies. We also address the problem of estimating the number of clusters. This is a challenging problem in cluster analysis, and several algorithms are available for it. We show that many of these existing algorithms have superior performance in high dimensions when they are constructed using MADD. We also construct a new estimator based on a penalized version of the Dunn index and prove its consistency in the HDLSS asymptotic regime. Several simulated and real data sets are analyzed to demonstrate the usefulness of MADD for cluster analysis of high dimensional data.

In this article, we develop and investigate a new classifier based on features extracted using spatial depth. Our construction is based on fitting a generalized additive model to posterior probabilities of different competing classes. To cope with possible multi-modal as well as non-elliptic nature of the population distribution, we also develop a localized version of spatial depth and use that with varying degrees of localization to build the classifier. Final classification is done by aggregating several posterior probability estimates, each of which is obtained using this localized spatial depth with a fixed scale of localization. The proposed classifier can be conveniently used even when the dimension of the data is larger than the sample size, and its good discriminatory power for such data has been established using theoretical as well as numerical results.

Parametric methods of classification assume specific parametric models for competing population densities (e.g., Gaussian population densities can lead to linear and quadratic discriminant analysis) and they work well when these model assumptions are valid. Violation in one or more of these parametric model assumptions often leads to a poor classifier. On the other hand, nonparametric classifiers (e.g., nearest-neighbor and kernel-based classifiers) are more flexible and free from parametric model assumptions. But, the statistical instability of these classifiers may lead to poor performance when we have small numbers of training sample observations. Nonparametric methods, however, do not use any parametric structure of population densities. Therefore, even when one has some additional information about population densities, that important information is not used to modify the nonparametric classification rule. This paper makes an attempt to overcome these limitations of parametric and nonparametric approaches and combines their strengths to develop some hybrid classification methods. We use some simulated examples and benchmark data sets to examine the performance of these hybrid discriminant analysis tools. Asymptotic results on their misclassification rates have been derived under appropriate regularity conditions.

In a classification problem, quite often the dimension of the measurement vector is large. Some of these measurements may not be important for separating the classes. Removal of these measurement variables not only reduces the computational cost but also leads to better understanding of class separability. There are some methods in the existing literature for reducing the dimensionality of a classification problem without losing much of the separability information. However, these dimension reduction procedures usually work well for linear classifiers. In the case where competing classes are not linearly separable, one has to look for ideal "features" which could be some transformations of one or more measurements. In this paper, we make an attempt to tackle both, the problems of dimension reduction and feature extraction, by considering a projection pursuit regression model. The single hidden layer perceptron model and some other popular models can be viewed as special cases of this model. An iterative algorithm based on backfitting is proposed to select the features dynamically, and cross-validation method is used to select the ideal number of features. We carry out an extensive simulation study to show the effectiveness of this fully automatic method.

Density estimates based on k-nearest neighbors have useful applications in nonparametric discriminant analysis. In classification problems, optimal values of k are usually estimated by minimizing the cross-validated misclassification rates. However, these cross-validation techniques allow only one value of k for each population density estimate, while in a classification problem, the optimum value of k for a class may also depend on its competing population densities. Further, it is computationally difficult to minimize the cross-validated error rate when there are several competing populations. Moreover, in addition to depending on the entire training data set, a good choice of k should also depend on the specific observation to be classified. Therefore, instead of using a single value of k for each population density estimate, it is more useful in practice to consider the results for multiple values of k to arrive at the final decision. This paper presents one such approach along with a graphical device, which gives more information about classification results for various choices of k and the related statistical uncertainties present there. The utility of this proposed methodology has been illustrated using some benchmark data sets.

Nearest neighbor classification is one of the simplest and most popular methods for statistical pattern recognition. A major issue in k-nearest neighbor classification is how to find an optimal value of the neighborhood parameter k. In practice, this value is generally estimated by the method of cross-validation. However, the ideal value of k in a classification problem not only depends on the entire data set, but also on the specific observation to be classified. Instead of using any single value of k, this paper studies results for a finite sequence of classifiers indexed by k. Along with the usual posterior probability estimates, a new measure, called the Bayesian measure of strength, is proposed and investigated in this paper as a measure of evidence for different classes. The results of these classifiers and their corresponding estimated misclassification probabilities are visually displayed using shaded strips. These plots provide an effective visualization of the evidence in favor of different classes when a given data point is to be classified. We also propose a simple weighted averaging technique that aggregates the results of different nearest neighbor classifiers to arrive at the final decision. Based on the analysis of several benchmark data sets, the proposed method is found to be better than using a single value of k.