Da varincia case we could also nd an est imat e for t

Info iconThis preview shows page 1. Sign up to view the full content.

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

Unformatted text preview: we could alsoassimno estimadorhe noise vé:iance estimador 2(3.5) ˆ • Sendo find a est imatρ Wr t para ε ar e o par amet er σ = ˆ n − 1 e( ρ∗) e( ρ∗), wher e e( ρ∗) = y − ρ∗W y − Z δ. ˆ T hi s suggest s an est i mdo termo estocástico Z ( I, onde: ) y . In t hi s at e f or δ of δ = ( Z Z ) − 1 é: n − ρ∗W . da variância case we could also find an est imat e for t he noise var iance par amet er σ2 = ˆ maximização δ. full log-likelihood da − 1 • Sendo assim a∗ ∗ ∗ ∗ ˆ n e( ρ ) e( ρ ), wher e e( ρ ) = y − ρ W y − Z para o modelo SAR => o Tirar e igualar a zero a primeira derivada com 2 respeito © 2009 by Taylor & Francis Group, LLC aos parâmetros β, σ e ρ. © 2009 by Taylor & Francis Group, LLC Resolver simultaneamente a condição de primeira ordem para todos os parâmetros. Forma alternativa de resolver o problema: loglikelihood function concentrated com respeito aos parâmetros β, σ2. o Calcula-se a solução de primeira ordem para β, σ...
View Full Document

This note was uploaded on 05/01/2013 for the course ECONOMIA 001 taught by Professor Farinha during the Fall '10 term at Universidade de Lisboa.

Ask a homework question - tutors are online