inst/shinyapps/rtlc/Prediction-template.md

In DimitriF/DLC: Deep Layer Chromatography

Prediction Templates

Those lines of code could be copy and paste in the editor tab to produce differents plots and print output

Variable importance for random forest

varImpPlot(model$finalModel,main='Variable importance for random forest')

Variable importances for random forest and plot of the first 3 variables

# First: extract the variable and reorder them
truc <- importance(model$finalModel)[order(importance(model$finalModel),decreasing=T),] 
names(truc) <- gsub('Ind.','',names(truc))

# Function that take the rank indice of importance to plot
f.plot <- function(vec = c(1,2)){ 
  plot(x=Ind[,names(truc)[vec[1]]],y=Ind[,names(truc)[vec[2]]],col=as.factor(Dep),
       xlab=names(truc)[vec[1]],ylab=names(truc)[vec[2]]
       )
}
par(mfrow=c(2,2)) # split the plot in 4 part
varImpPlot(model$finalModel)
# Change the number in the funciton to plot a different set of variable
f.plot(c(1,2))
f.plot(c(1,3))
f.plot(c(2,3))

Score plot for pls

data <- as.data.frame(scores(model$finalModel)[,1:2])
colnames(data) <- c('Comp1','Comp2')
Dep <- Dep[Train.partition()]
print(qplot(Comp1,Comp2,data=data,col=Dep))

DimitriF/DLC documentation built on Oct. 14, 2020, 4:33 p.m.