search

jaredsmurray / bcf

Bayesian Causal Forests
40 stars 20 forks source link

Is there any reason to comment out 'mu' element output?? #3

Closed tmitanitky closed 4 years ago

tmitanitky commented 5 years ago

Thank you for a great paper and implementation.

I got trouble in visualizing tau and mu separately.

I couldn't get mu output from bcf_fit = bcf(y, z, x, x, pihat, nburn=2000, nsim=2000) to find the line 284 in bcf is commented out. Is there any reason for the comment out line 284 in bcf.R?

https://rdrr.io/cran/bcf/man/bcf.html tells bcf return mu, but the explanation for mu would be "nsim by n matrix of posterior samples of prognostic scores".

https://github.com/jaredsmurray/bcf/blob/cb5dcfc9b611dcf721ed09bdca88ff6c152946e1/R/bcf.R#L282-L289

jaredsmurray commented 4 years ago

mu as commented out would be on a strange scale. The prognostic score E(Y|z=0, x=0) is given by yhat if the observation has z=0 or yhat - tau if the observation has z=1.