A quick review for the previous parts
1
2Recall the assumptions in regression
⏹
Population model ⏹Assumption A.1-A.5
μββ++=x y 10
(Linearity)i i i u x y ++=10ββ,
0)|(=x u
E j i u u E x u Var j i ≠==,0)()|(2
σ
),0.(..~|2
σd i i x
u )
,0(...~|2σN d i i x
u (for hypothesis testing)(A.2) and (A.3) imply (zero conditional mean)x
x y E 10)|(ββ+=From A.2
3
Criteria for choosing good (ideal) estimatorvariable used in lambda
1.Unbiasedness(无偏)
2.Consistency(一致)
3.Efficiency (smallest variance)(有效)
If the regression model satisfies assumption A.1 to A.4, In the finite sample , the estimators from OLS are BLUE (Gauss-Markov theorem)
1.Linear
2.Unbiased
3.Smallest variance among all the linear unbiased estimators In the large sample
consistent
regression assumptions ⏹Assumption A.3
of OLS

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。