说明：  线性预测分析是最有效的语音分析技术之一，在语音编码、语音合成、语音识别和说话 人识别等语音处理领域中得到了广泛的应用。语音线性预测的基本思想是：一个语音信号的 抽样值可以用过去若干个取样值的线性组合来逼近。通过使实际语音抽样值与线性预测抽样 值的均方误差达到最小，可以确定唯一的一组线性预测系数。
(Linear predictive analysis is one of the most effective voice analysis technology has been widely applied in the field of speech processing speech coding, speech synthesis, speech recognition and speaker recognition and the like. The basic idea of ​ ​ linear prediction of speech is: a voice signal sample values ​ ​ can be approximated by a linear combination of a plurality of sample values ​ ​ of the past. By making the actual speech sample mean square error of the value of the smallest linear prediction sample values, to determine a unique set of linear prediction coefficients.)