In this paper, a simple three_layers BP network was used for the early prediction of the amount of product, based on the difference in prediction errors between normal and abnormal charges and other accessorial information, such as profit function.

The use of simplified model and prediction error correction improves conventional control algorithm.

In order to better improve dynamical seasonal prediction by utilizing statistical experiences, the problem of capturing statistical experiences associated with climate model is discussed from the point of view of studying the impacts of physical predictors in climate system on model prediction errors.

Analysis on forecast error of extratropical cyclone in winter American and Japanese radiofacsimile weather charts;

based algorithm with the correction of a forecast error has been put forward to fulfill relevant control functions.

Using the equatorial Pacific sea surface temperature anomaly(SSTA) forecasts of the Zebiak-Cane model,the SSTA prediction skill for neutral years and the growing/decaying phase of El Nio/La Nia events are investigated by analyzing the behavior of forecast error growth.

It is proved to have the same asymptotic statistical properties as the prediction error method(PEM).

The main contribution of this paper is that an error prediction based method of control system comprehensive performance assessment is proposed, which is eligible for both stochastic and deterministic performance evaluations.

An extrapolative formula for calculation of minimum mean square error forecast Xt(l) is obtained and a recursive formula for computation of the conditional variance for forecasting error Xt+l-Xt(l) is derived.