This paper presents efficient artificial neural network based linear p
redictive coding. Speech recognition is fundamentally pattern classifi
cation task. The objective is to input pattern, the speech and classif
y a sequence of patterns. The linear predictive coefficients of the sl
owly varying speech signals are stored. A feedforward network is deter
mined by linear predictive coefficients. A three layered feedforward n
etwork was used with back propagation as the training algorithm. The n
etwork and learning techniques are proved for their correctness and ap
plied to the problem of speech recognition. The suggested novel scheme
yields results in finite accuracy and recognition performance.