Approach to Piecewise-Linear Classification in a Multi-dimensional Space of Features Based on Plane Visualization

Abstract

Information technology allowing to implement problems of classification, clustering, researching of topology of the information component of the data is proposed. The multidimensional feature space is reduced to the visual presentation space to determine the information content of the data. Optimized reduction of the space dimension to two-dimensional one applying multidimensional scaling methods has been used. The use of the piecewise linear constraints allows us to implement projecting into original multidimensional feature space. Visual construction of restrictive separators makes possible consideration of tolerance fields of changing of features parameters, measure separation of classes, nonlinearity of data grouping. Thus, restrictive areas of hyperspace for the necessary categories of classes are formed. At the same time, visualization of the classification processes in hyperspace has been provided. Information technology is the multidimensional space projecting into visual (two-dimensional) space, constructing of piecewise linear constraints of studied areas, subsequent constraints projecting into multidimensional space. Thus, the information technology enables to synthesize separating hyperplanes limiting categories of classes in multidimensional space. The technology application successive stages and example for fingerspelling alphabet recognition have been described.