Update priors not from training-samples as it is given, but from distribution of our algo (ie RF)
tsundokum opened this issue · comments
Marat Zaynutdinov commented
Marat Zaynutdinov commented
Update priors not from training-samples as it is given in a raw file, but from distribution of our algo (ie RF) on that file. Very possible, that our algo alters distribution of classes in the raw file.
PGK commented
Look cryptic to me. What raw file do you mean?