prediction error
1.MINIMUM PREDICTION ERROR ADAPTIVE CONTROLLER BASED ON NEURAL NETWORK MODEL
4.In the method, the criterion of final prediction error (FPE) is employed to determine the embedding dimension of samples.
5.The absolute value of prediction error is within 0～0.7 m,and about 95% of the values come closely to 0～0.3 m.
6.Three types of adaptive filters, the adaptive prediction error filter (APEF), the two-side tapped adaptive transversal filter(ATF) and the adaptive lattice filter (ALF), are considered.
7.Based on the theory of recursive prediction error method,a new identification algorithm is developedfor the multi-input multi-output ARMAX model A(z)y_(?) =B(z)u_(?) +C(z)ε_(?)
8.The prediction error percentage of the 52000t all-purpose cargo ship compared with the model test result is 21.8%, 300000t VLCC compared with the measured value is merely 3.64%.
9.The optimal interval of case dissimilarity threshold R for CBR was determined by a CFW forecast test,and the best mean prediction error rate of the ill plant rate and the ill leaf rate were 7.4% and 9.3% respectively.
10.Closed form expressions are der-ived for estimates using the autoregressive (AR) prediction error filter approach,as well as using periodogram with Bartlett window, and the maximum likelihood(ML) method. These expressions are useful in the study of resolving closelyspaced sinusoidal signals.
