Consider the problem of regression through the origin with n

Consider the problem of regression through the origin (with normal error terms, all with the same variance) where the X values in the training data set take on the values for some number n. Let beta denote the one-dimensional regression coefficient, and let sigma 2 denote the variability of the outcome around it\'s conditional expectation. Suppose now that we assume the has a normal prior with variance lambda -1 and expectation zero. Assume that Xnew will follow a uniform distribution on the unit interval.

Solution