Working with Soilspec data (part 10)

 Now it is time for regressions and prediction for all the parameters using the selected spectra with the "puchwain"  function, and for test, the non selected ones.

We develop the regressions with Caret using PLS and Cross Validation (the model choose 5 terms).

model_clay_snvdt <- train(y_sel_clay ~.,data=trainDataClay,  

                          method = "pls", scale = TRUE,

                          trControl = trainControl("cv", number = 10),

                          tuneLength = 20)

We can plot the predictions of the Training Set (selected samples) with the predictions of the Test Set (non selected Samples) for every parameter. I do in this case for Clay :