[Gretl-users] Re-run OLS regression with only statistically important and positive variables via scripting