G.V. Malshakov - Senior Lecturer, Department 403, MAI. E-mail: malshakov@mail.ru

Errors of skipping detection and false detection, arising from identification of entities of applied software, are investigated on the basis of the frequency analysis of data. Dependences of a correlation error on the sample size for the bipolar transistor with isolated shutter (BTIS), quantities of errors of «the miss of essence» arising for 178 entities participating in the research from the sample size, an error of correlation and quantity of lexemes in the alphabet from frequency restriction of the alphabet, autocorrelation from a minimum quantity-admissible of symbols in lexemes for various selections (in % from base data) are shown; correlation from a minimum quantity-admissible of symbols in lexemes, autocorrelation of BTIS and correlation of the field transistor from a minimum quantity-admissible of symbols in lexemes for various selection, false and correct detection entity from quantity of nonzero meetings of lexemes of the alphabet are studied. The condition on quantity of nonzero values of a meeting of lexemes of the alphabet, providing optimum parity between quantity of errors of false detection and an opportunity of correct detection is established. The received results have practical value for improvement of quality of identification of entities of applied software, carried out on the basis of the frequency analysis of data.