BUILDING A MODEL FOR CLASSIFICATION OF GRAIN BULK PRODUCTS BY DIFFUSE REFLECTION SPECTRA IN THE NEAR INFRARED REGION ON THE EXAMPLE OF LOGISTIC REGRESSION

Abstract

The possibility of determining a grain corps having a different grinding by the diffuse reflection spectrum was experimentally confirmed. Combinations of optical densities and their second derivatives for the wavelengths of 1200, 1422, 1778, 1916, and 2114 nm were used as features describing the diffuse reflection spectra of wheat and oats of different milling and humidity in the near infrared range. Using the example of logistic regression, 20 classification models based on two features were constructed: 10 models for optical density and 10 models for the second derivative of the optical density corresponding to the selected wavelengths. The best classification results were obtained by an algorithm that used the values of the second derivative of optical density at 1778 and 2114 nm as features.