We show a simple example in which we fit PheNorm to a simulated dataset.
Load the simulated data
fit.dat <- read.csv("https://raw.githubusercontent.com/celehs/PheNorm/master/data-raw/data.csv") str(fit.dat) #> 'data.frame': 500 obs. of 4 variables: #> $ X1 : num -1.38 0.995 0.524 1.48 1.344 ... #> $ X2 : num -1.73 -0.624 1.757 -0.563 0.296 ... #> $ ICD: num 0.693 1.386 1.386 1.386 1.792 ... #> $ utl: num 3.78 3.91 3.83 3.87 3.64 ...
Apply the PheNorm algorithm
set.seed(1234) fit.phenorm <- PheNorm.Prob( nm.logS.ori = "ICD", nm.utl = "utl", dat = fit.dat, nm.X = NULL, corrupt.rate = 0.3, train.size = nrow(fit.dat) )
str(fit.phenorm) #> List of 2 #> $ probs: num [1:500] 0.466 0.538 0.546 0.542 0.609 ... #> $ betas: NULL