Entropy - Section ARE213 Entropy Econometrics Estimation Procedure by Golan Judge and Miller 1996 May 2006 Hendrik Wolff Short Review on MaxEntropy

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Section ARE213 Entropy Econometrics Estimation Procedure by Golan, Judge and Miller, 1996 May 2006 Hendrik Wolff Short Review on MaxEntropy 1948 Claude Shannon: Information Entropy: H( p ) = - ∑ j=1 p j ln p j Example of Dice: 6 Support Points z j : z 1 , z 2 ,..., z 6 Max p H( p ) s.t. 3.5 = ∑ j=1...6 p j z j 1.0 = ∑ j=1 p j Solution: p j =1/6: I.e. Uniform distribution of the discrete PDF f(z) How does f(z) look like, if we don‘t observe the theoretical mean of 3.5 ? Solution to the ME-Problem Max p H( p ) = Max p [- ∑ j=1 p j ln p j | y = ∑ j=1...6 p j z j , 1.0 = ∑ j=1 p j ] Lagrange: L = - ∑ j=1 p j ln p j +λ (y - Z p ) + θ(1.0 - p´1 ) δL/δ p = - ln p - 1- Z´ λ - θ = 0 δL/δ λ = y- Z p = 0 δL/δ θ = 1- p´1 = 0 --> p k = exp(-Z k ´λ) / Ω k (λ) with Ω(λ) = Σ i=1...6 exp(-Z k ´λ) No analytical solution (parallels Logit) Newton worsk since H is globally concave Normalized Entropy Measuring the Information Content: „Importance of the contribution of each piece of data in reducing...
View Full Document

This note was uploaded on 08/01/2008 for the course ARE 213 taught by Professor Imbens during the Spring '06 term at University of California, Berkeley.

Page1 / 13

Entropy - Section ARE213 Entropy Econometrics Estimation Procedure by Golan Judge and Miller 1996 May 2006 Hendrik Wolff Short Review on MaxEntropy

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online