gretl <-> R interaction
by Riccardo (Jack) Lucchetti
Hi everyone,
last week, Allin and I worked on better integration with R, and it would
be very helpful if some of you guys could test a new feature, which is
available in the current CVS version and/or the windows snapshot: you can
now edit R sources and have R execute them, with the data in your current
gretl session auto-loaded for you.
An example will probably clarify:
1) load the "data4-1.gdt" example file
2) run an OLS regression of "price" on the other 3 variables: you should
get something like
const 129.062 88.3033 1.462 0.17456
sqft 0.154800 0.0319404 4.847 0.00067 ***
bedrms -21.5875 27.0293 -0.799 0.44304
baths -12.1928 43.2500 -0.282 0.78376
3) now open Files > Script files > New and choose "R script"
4) an edit window should open: write the following R commands:
mod1 <- lm(price ~ sqft+bedrms+baths)
summary(mod1)
5) click on the gears icon and choose "Non interactive"
6) after a few moments, an output window should open with the results,
among which:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 129.06163 88.30326 1.462 0.174559
sqft 0.15480 0.03194 4.847 0.000675 ***
bedrms -21.58752 27.02933 -0.799 0.443037
baths -12.19276 43.25000 -0.282 0.783758
(Hey, the same numbers! Amazing!)
IMHO, this mechanism (if working properly) will be very beneficial to
gretl _and_ R users. However, it is not very well tested, and anything can
happen (crashes, data losses, kittens being eaten, your house on fire).
So, I'd like to invite anyone who has even a marginal interest in R to
give this new feature a good run, so we can debug it properly before the
next release.
Thanks.
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti
