Likelihood Talk v2

Consider again, the mean and width of a Gaussian. ( ) 2 0 * 2 i * i i 0 0 * N 1 N x N 1 N 1 x N 1 N 1 2 ms m--= = å å å å Note that the MLE of the mean is unbiased, but for the width is not (although it is consistent). However, is unbiased. In this case, we could find the bias analytically - in most cases we must look for it numerically. s ... ................
................