Closed bi0m3trics closed 5 years ago
Sorry about that, I added a feature request for a classification instance and in mucking with the code broke the regression sampling. It is fixed now. Please install the development version off of GitHub as this will not be on CRAN for a bit.
Best, Jeff
Jeffrey S. Evans, Ph.D., | Senior Landscape Ecologist / Biometrician The Nature Conservancy | Global Lands Science Team Visiting Professor | University of Wyoming | Zoology & Physology Laramie, WY | jeffrey_evans@tnc.orgmailto:jeffrey_evans@tnc.org | (970) 672-6766<tel:(970)%20672-6766>
From: Andrew Sánchez Meador notifications@github.com Sent: Friday, February 1, 2019 7:50 AM To: jeffreyevans/rfUtilities rfUtilities@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: [jeffreyevans/rfUtilities] Error in rf.crossValidation for rfUtilities 2.1-4 (#4)
Hi Jeffrey - it looks like the Dec 14th 2018 commithttps://github.com/jeffreyevans/rfUtilities/commit/a511ee910e8b32eb87e8f21399599071b0e68958#diff-70b42b94250cb084705b7fde91693a41 introduced an error on line 142, where sample.sizes is being called but it appears to have not been initiated in the regression cross-validation code (...this is not the case in the classification cross-validation code on lines 184-187). The following code (running randomForest 4.6-14 and rfUtilities 2.1-4 in R 3.5.2) from the documentation:
data(airquality)
airquality <- na.omit(airquality)
rf.mdl <- randomForest(y=airquality[,"Ozone"], x=airquality[,2:4])
( rf.cv <- rf.crossValidation(rf.mdl, airquality[,2:4], p=0.10, n=99, ntree=501) )
produces the following error:
Error in sample.int(length(x), size, replace, prob) :
object 'sample.sizes' not found
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/jeffreyevans/rfUtilities/issues/4, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AJaj_3nn84UxQlYX4DQZUDvfCwVkZ-WMks5vJFQdgaJpZM4aedLf.
Hi Jeffrey - it looks like the Dec 14th 2018 commit introduced an error on line 142, where
sample.sizes
is being called but it appears to have not been initiated in the regression cross-validation code (...this is not the case in the classification cross-validation code on lines 184-187). The following code (runningrandomForest
4.6-14 andrfUtilities
2.1-4 inR
3.5.2) from the documentation:produces the following error: