search

kogalur / randomForestSRC

DOCUMENTATION:
https://www.randomforestsrc.org/
GNU General Public License v3.0
113 stars 18 forks source link

survival analysis with 300 lipids #247

Closed MoMaz123 closed 2 years ago

MoMaz123 commented 2 years ago

Hi, I want to run survival analysis with 300 lipids to find out the VIP variables, should I just use normal survival code? MMODEL <- rfsrc(Surv(time, status) ~ ., data = MYDTATA, ntree = 100, block.size = 1)

Many thanks

ishwaran commented 2 years ago

For anti-VIMP (default) MMODEL <- rfsrc(Surv(time, status) ~ ., data = MYDTATA, importance = TRUE)

For permute VIMP MMODEL <- rfsrc(Surv(time, status) ~ ., data = MYDTATA, importance = "permute")

See also: https://www.randomforestsrc.org/reference/vimp.rfsrc.html and https://www.randomforestsrc.org/articles/vimp.html

MoMaz123 commented 2 years ago

Many thanks, what is anti-vimp please,

ishwaran commented 2 years ago

See the help file for rfsrc and read Section 6 on VIMP. Most of your questions are already explained in the help file, with further references for reading, which you should first consult before posting here. Thanks.

https://www.randomforestsrc.org/reference/rfsrc.html