| Popis: |
The eigenvector approach in matrix theory is used to study a stochastic differential equation whose resolution corresponds with that of the least mean square (LMS) algorithm, based on Ljung's (1977) stochastic theory. A conclusion comes from our study that the trace of the weight adjustment of the algorithm is, in statistics, the same as the transitional process of an RC circuit of first order. As a result, the device designed with the algorithm has the same properties as a parallel filtering array which consists of some RC circuits of first order has. A relationship between the cut-off frequency of an adaptive system and the eigenvalues (power components) as well as band width of a signal is highlighted. |
