ubject: [R] Linear Regression with Linear Equality Constraint
>
>Please advice on the package I should use to run a linear regression
>model
>(weighted least squared) with linear equality constraint. I initially
>tried
>"constrOptim" but it turned out that it only suppo
On Oct 31, 2011, at 5:12 PM, Comcast wrote:
>
>
> On Oct 31, 2011, at 3:31 PM, Bert Gunter wrote:
>
>> Well, if I understand the question correctly (following the posting guide
>> would have spared guessing, as usual), forget packages -- nothing more than
>> elementary algebra is needed.
>>
On Oct 31, 2011, at 3:31 PM, Bert Gunter wrote:
> Well, if I understand the question correctly (following the posting guide
> would have spared guessing, as usual), forget packages -- nothing more than
> elementary algebra is needed.
>
> e.g.
>
> lm(y ~ x1 + x2 + x3) subject to the constrain
Well, if I understand the question correctly (following the posting guide
would have spared guessing, as usual), forget packages -- nothing more than
elementary algebra is needed.
e.g.
lm(y ~ x1 + x2 + x3) subject to the constraint beta_2 = beta_3, the
coefficients of x2 and x3, is the same as
I believe the package systemfit can help you with that. Haven't tried it
myself, but give it a go.
Regards,
Kristian
2011/10/31 JW
> Please advice on the package I should use to run a linear regression model
> (weighted least squared) with linear equality constraint. I initially
> tried
> "co
Please advice on the package I should use to run a linear regression model
(weighted least squared) with linear equality constraint. I initially tried
"constrOptim" but it turned out that it only supported inequality linear
constraint. Thank you very much in advance.
Cheers,
Jon
--
View this m
6 matches
Mail list logo
