About the full prediction approximation by a lot of partial predictions in case of incomplete data

In this article, we are talking about the random subspaces method in forecasting under condition of incomplete data and estimation a full forecast based on set partial predictions. Centered predictions considered without loss generality. According to statistical model, off-diagonal elements correlation matrix with known mathematical expectation variance. case matrix, analytical expressions obtained for determinant minors. Based these results, class more accurate estimates prediction is constructed, which differ from mean by multipliers that depend parameters The results modeling practical biogeographic presented.