ADAPTS Vignette #5
Repeated Leave-Half-Out Cross-Validation for Robust Accuracy Estimates
Xinyi Yan
8/21/2020
Install ADAPTSdata4 using the code:
install.packages(‘devtools’)
library(devtools)
devtools::install_github(‘yxinyi2/ADAPTSdata4’)
Construct all the signature matrices one time
This prostate data set comes from Henry, Gervaise H., et al. “A cellular anatomy of the normal adult human prostate and prostatic urethra.” Cell reports 25.12 (2018): 3530-3542.
Half of the data are held out as test set by setting ‘testOnHalf’ to TRUE.
Result tables are for original cell types and clustered cell types relatively.
scHenry<-as.matrix(log(ADAPTSdata4::scHenry+1))
resList<-testAllSigMatrices(exprData=scHenry, randomize = TRUE, skipShrink=FALSE, proportional=FALSE,handMetaCluster=NULL,testOnHalf=TRUE)
