On one method of saving memory when classifying texts

The article investigates the method of memory saving in tasks of classification of texts by searching for matching parts of linear polynomials. The algorithm for finding matching parts in linear polynomials with integer coefficients is given at the beginning. This algorithm makes it possible to calculate systems of linear polynomials with integer coefficients more quickly and use less memory for their storage. The algorithm is then used to find the matching parts of the linear polynomials that arise when classifying texts using the Bayesian classifier. We provide computational experiments that show memory saving. (In Russian).