Replacing R Squared with AIC

[Giles German]

Dear All,

I have the following equation which is a standard time series analysis where I am trying to replace the highest R-Squared with the lowest AIC....

Can you comment on how I replace the code?

Please not all the variables have been loaded in the loop. Happy to give details as required.

foreach var of local lvif {
  `idf' `var'
  if( e(r2) > `r2'){
    local r2=e(r2)
    local iidf "`idf' `var'"
    }
}

My effort below...

scalar lowest_aic = c(maxdouble)
foreach var of local lvif {
  `idf' `var'
  estat ic
  matrix S =r(S)
  scalar aic = S[1, 5]
  if aic < lowest_aic {
    local result "`idf'`var'"
    scalar lowest_aic = aic
  }
}

Giles

[Clyde Schechter]

And your question is....?