iv Cov bracketleftbig β Θ j X j bracketrightbig E braceleftbig Cov

# Iv cov bracketleftbig β θ j x j bracketrightbig e

• 10

This preview shows page 5 - 8 out of 10 pages.

(iv) Cov bracketleftbig β j ) , X j bracketrightbig = E braceleftbig Cov bracketleftbig β j ) , X j | Θ j bracketrightbigbracerightbig + Cov braceleftbig E bracketleftbig β j ) | Θ j bracketrightbig , E [ X j | Θ j ] bracerightbig = E (0) + Cov bracketleftbig β j ) , Y β j ) bracketrightbig = A Y prime . (v) Cov bracketleftbig ˆ β j , β j ) bracketrightbig = Cov bracketleftbig ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j X j , β j ) bracketrightbig = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j Cov bracketleftbig X j , β j ) bracketrightbig = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j ( Cov bracketleftbig β j ) , X j bracketrightbig) prime = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j ( Y prime ) prime A prime = A .

Subscribe to view the full document.

64 CHAPTER 6. CREDIBILITY REGRESSION MODELS (vi) Cov[ ˆ β j , X j ] = Cov[( Y prime V - 1 j Y ) - 1 Y prime V - 1 j X j , X j ] = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j Cov( X j ) = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j ( s 2 V j + Y A Y prime ) = s 2 ( Y prime V - 1 j Y ) - 1 Y prime + A Y prime = s 2 U j Y prime + A Y prime . (vii) Cov bracketleftbig ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j X j bracketrightbig = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j Cov( X j ) V - 1 j Y bracketleftbig ( Y prime V - 1 j Y ) - 1 ] prime = ( Y prime V - 1 j Y ) - 1 Y prime V - 1 j ( s 2 V j + Y A Y prime ) V - 1 j Y ( Y prime V - 1 j Y ) - 1 = s 2 ( Y prime V - 1 j Y ) - 1 + A = s 2 ( Y prime V - 1 j Y ) - 1 + A = s 2 U j + A . a50 Theorem 6.2. The non–homogeneous linear regression credibility estimator B a j of β j ) that minimises E braceleftbigbracketleftbig β j ) - B a j bracketrightbig prime Σ j bracketleftbig β j ) - B a j bracketrightbigbracerightbig is given by B a j = Z j ˆ β j + ( I - Z j ) b , where Z j = A ( A + s 2 U j ) - 1 = Cov bracketleftbig β j ) bracketrightbig bracketleftbig Cov( ˆ β j ) bracketrightbig - 1 . (6.5) Hence the non–homogeneous linear regression credibility estimator M a j of μ j ) is given by M a j = Y B a j . *Proof : Let g j = c 0 + c 1 X 0 j be an arbitrary vector that is a non–homogeneous linear combination of the X j and epsilon1 0. Then for any positive definite matrix Σ j , consider d j ( epsilon1 ) = E braceleftbigbracketleftbig β j ) - B a j - epsilon1g j bracketrightbig prime Σ j bracketleftbig β j ) - B a j - epsilon1 g j bracketrightbigbracerightbig .
6.2. HACHEMEISTER’S REGRESSION MODEL 65 The theorem holds if d prime j (0) = 0 for any g j . Now, by Lemma 6.1, d prime j ( epsilon1 ) = - 2 E braceleftbig g prime j Σ j bracketleftbig β j ) - B a j bracketrightbigbracerightbig + 2 epsilon1 E bracketleftbig g prime j Σ j g j bracketrightbig = - 2 E braceleftbig g prime j Σ j bracketleftbig β j ) - B a j - epsilon1g j bracketrightbigbracerightbig . (6.6) Then define the following reduced variables β 0 j ) = β j ) - b , ˆ β 0 j = ˆ β j - b , X 0 j = X j - Y b j , so that our normal equation (6.6) at epsilon1 = 0 reduces to E braceleftbig g prime j Σ j bracketleftbig β 0 j ) - Z j ˆ β 0 j bracketrightbigbracerightbig = 0 . Now since g j = c 0 + c 1 X 0 j , equivalently we need to prove that E braceleftbigbracketleftbig c prime 0 + ( X 0 j ) prime c prime 1 bracketrightbig Σ j bracketleftbig β 0 j ) - Z j ˆ β 0 j bracketrightbigbracerightbig = 0 .

Subscribe to view the full document.

• Fall '09
• Dr.D.Dryanov

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern

Ask Expert Tutors You can ask 0 bonus questions You can ask 0 questions (0 expire soon) You can ask 0 questions (will expire )
Answers in as fast as 15 minutes