М-Модели алгоритмов. Емкость и колмогоровская сложность класса М-полиномов.
The problems with elements bounded by a bit array are extracted in a special class of learning by precedents problems. A notion of M-models of learning algorithm is introduced. The Kolmogorov complexity and the VCD of M-polynomials and M-polynomials of Zhegalkin with k-component are estimated. The notions of complexity and a degree of compression by algorithms of M-models for training samples are introduced.