[R] systemfit: data

[Martin Ivanov]

Dear R users,
I have a question on the function systemfit from the systemfit package. This function returns a list of the class systemfit, which besides all other information, contains the data frame of the whole system, called “data”. I have noticed the following issue.
Suppose that I have to estimate a single equation (the simplest case), for example:
x ~ y + z

so I call systemfit like that:

a <- systemfit(method=”OLS”,eqns=formula(“x~y+z”),data=data.frame(cbind(x,y,z)))

Then, a$data contains the variables x,y and z in its columns. So far so good. However, if I do not want an intercept term to be evaluated and I call systemfit like that:
a <- systemfit(method=”OLS”,eqns=formula(“x ~ -1 + y + z”),data=data.frame(cbind(x,y,z)))

then a$data contains only x and z in its columns. The regressor y is missing.
The same issue is valid for systems of equations: if an intercept term is not allowed, “data” contains everything it should with the exception of the first variables on the right hand side of the equations. If an intercept term is allowed, then “data” does contain the data frame of the whole system.

If I do something wrong, please tell me. 
I am looking forward to your reply.
Regards,
Martin

-----------------------------------------------------------------
Всички учебници и помагала - отстъпка+доставка WWW.MOBILIS.BG