For the linear regression model

Y=XB+e

(B= beta and e = epsilon)

ei~N(0,s^2) (s = sigma)

i = 1,...,n are independant, this model does not contain an intercept form

(E = sum of , ^B = beta hat)

i found that ^B = Ei xiyi / Ei xi^2

now i need to show that ^B is an unbiased estimator of B and that its variance is s^2/ Ei xi^2

jus wondering what the proof is for a linear regression model that the expected or estimated value of ^B =B in this circumstance and how to prove that var(^B)=s^2(XtX)^-1

ive been working in matrix form, t=transpose as im sure u might have guessed