ABSTRACT
This chapter discusses the prediction of the least mean-square error (LMSE) and surveys the techniques for predicting under various circumstances. It describes the multivariate LMSE prediction by using normal equations. A predictor is said to be unbiased if it has the “right” expected value. There is no mechanism in the linear LMSE predictor that enforces this condition. The chapter discusses interpolation, smoothing, extrapolation, and back-prediction. It also explains the use of Wiener filter in LMSE prediction.
