Мінімаксні середньоквадратичні оцінки тренду в задачах регресії

This paper describes an approach to the linear minimax estimation of the generalized polinom with unknown partially unbounded parameters. Linear minimax estimations are constructed on the basis of measurements with random noise. We introduce notions of the upper and lower linear minimax estimations. The sufficient conditions for the minimax estimation existence are formulated. Numerical simulation that illustrates these results is presented.