This article is cited in 1 paper

Administration of technological processes

Identification of statistical process models under incomplete information

L. A. Kuznetsov, A. M. Korneev, M. G. Zhuravlyova

Lipetsk State Technical University

Abstract: The possibility of EM algorithm strategy application for building multi ple regression models under incomplete process data is examined. The models based on complete data and the data were the missing points were replaced with unconditional and conditional means are compared. An enhanced model describing the relationship between a mechanical characteristic of flat and a set of process parameters is obtained with the help of EM algorithm.

UDC: 681.51.015