Abstract :
The application of series representations for random processes to problems of causal least-mean-square estimation on infinite intervals has met with some difficulty in the past. This correspondence brings to light two of the more outstanding difficulties in relation to the contribution of Cambanis, which appears in this issue.
