Consider the multiple regression model y=xβ+εlon. The Gauss-Markov conditions hold: E[εlon]=0 and var[εlon]=σ²I. In addition, εloni∼N(0,σ). We have shown that the estimator of (β) is given by (β)=(x'x)⁻¹x'y.
a. Use properties of the trace to find E[y'(I-H)y].