THE FORMATION CLASSES OF OBJECTS BY THE METHOD OF DISCRIMINANT ANALYSIS

Abstract

Summary.The paper provides the method of discriminant analysis as a modern tool for the classification of objects by the example of flour production. Discriminant analysis is a statistical technique that allows us to study the differences between two or more groups of objects on multiple variables simultaneously and provides the ability to classify objects according to the principle of maximum similarity. Content of discriminant analysis is the development and study of statistical methods to examine the differences between two or more groups of objects on multiple variables simultaneously with the dominant line. In discriminant analysis, in contrast to the cluster, there is a training set, which is known what classes are objects. The training set is obtained rules, which further allow you to determine what class are new objects. Built discriminant functions, graphs of distribution of objects on quality classes, graphically presents classification methodology. During the performance was formed database consisting of 595 analyzes characterizing the quality of flour by 15 characters. Each assay described chemical parameters (mass fraction of protein mass fraction of ash, the mass fraction of fat, fiber content and water-soluble carbohydrates) and organoleptic quality of flour (moisture content, titratable acidity and active, and the mass fraction of gluten quality, taste, smell, and the crunch etc.). Classification accuracy of the method of discriminant analysis was 576 (98.02%).