[R-br] Res: regressão multivariada

[Bernardo Rangel Tura]

On Tue, 2011-04-19 at 17:01 -0700, Ayuni Sena wrote:
> Muito obrigada Meninos!
> Bem lembrado Benilton....o que eu quero é fazer o ajuste simultaneo da
> matrix Y em função da matriz X 
> agradeço a atenção de voces.
> Ayuni


É por isso que sugeri o systemfit, vejam o exemplo com apenas um
modificação

Codigo:

require('systemfit')
data("Kmenta")
attach(Kmenta)
eqDemand <- consump ~ price + income
eqSupply <- consump ~ price + farmPrice + trend
eqSystem <- list(demand = eqDemand, supply = eqSupply)
fitols <- systemfit(eqSystem)
summary(fitols)

resposta:

systemfit results 
method: OLS 

        N DF     SSR detRCov   OLS-R2 McElroy-R2
system 40 33 155.883 4.43485 0.709298   0.557559

        N DF     SSR     MSE    RMSE       R2   Adj R2
demand 20 17 63.3317 3.72539 1.93013 0.763789 0.735999
supply 20 16 92.5511 5.78444 2.40509 0.654807 0.590084

The covariance matrix of the residuals
        demand  supply
demand 3.72539 4.13696
supply 4.13696 5.78444

The correlations of the residuals
         demand   supply
demand 1.000000 0.891179
supply 0.891179 1.000000


OLS estimates for 'demand' (equation 1)
Model Formula: consump ~ price + income

              Estimate Std. Error  t value   Pr(>|t|)    
(Intercept) 99.8954229  7.5193621 13.28509 2.0906e-10 ***
price       -0.3162988  0.0906774 -3.48818  0.0028153 ** 
income       0.3346356  0.0454218  7.36729 1.0999e-06 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

Residual standard error: 1.930127 on 17 degrees of freedom
Number of observations: 20 Degrees of Freedom: 17 
SSR: 63.33165 MSE: 3.725391 Root MSE: 1.930127 
Multiple R-Squared: 0.763789 Adjusted R-Squared: 0.735999 


OLS estimates for 'supply' (equation 2)
Model Formula: consump ~ price + farmPrice + trend

              Estimate Std. Error t value   Pr(>|t|)    
(Intercept) 58.2754312 11.4629099 5.08383 0.00011056 ***
price        0.1603666  0.0948839 1.69013 0.11038810    
farmPrice    0.2481333  0.0461879 5.37226 6.2274e-05 ***
trend        0.2483023  0.0975178 2.54623 0.02156713 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

Residual standard error: 2.405087 on 16 degrees of freedom
Number of observations: 20 Degrees of Freedom: 16 
SSR: 92.551058 MSE: 5.784441 Root MSE: 2.405087 
Multiple R-Squared: 0.654807 Adjusted R-Squared: 0.590084 

fim da reposta

Obs você pode usar outras formas de ajuste com SURE por exemplo


-- 
[]s
Tura