After coarse frequency synchronization, frequency offset compensation is carried out using variable step size least mean square algorithm (VLMS).
经过频率粗同步后，以变步长的最小均方算法(VLMS)进行频率补偿。
The connection weights between hidden layer and output layer are got by Least Mean Square Algorithm.
径向基函数神经网络的隐含层到输出层的线性连接权值，则是由最小二乘法来计算得到的。
It is shown in this paper that the Least-Mean-Square algorithm cannot minimize the mean square error, and that it is actually a kind of weighted Least-Square algorithm.
本文指出最小均方算法并不能使均方误差最小，证明了最小均方算法实际上是一种加权最小二乘算法。
NLMS (normalized least mean square) algorithm is widely used in echo canceller. However, due to high correlation of speech signals, the performance of echo canceller based on NLMS is depressed.
NLMS算法是回声消除器中最常用的算法之一，然而语音信号的强相关性使NLMS(归一化最小均方)算法的收敛速度变慢。
NLMS (normalized least mean square) algorithm is widely used in echo canceller. However, due to high correlation of speech signals, the performance of echo canceller based on NLMS is depressed.
NLMS算法是回声消除器中最常用的算法之一，然而语音信号的强相关性使NLMS(归一化最小均方)算法的收敛速度变慢。
应用推荐