hs_RF: Train random forest model on hyperspec object

In CMET-UGent/MicroRaman: Phenotyping of microbial Raman spectroscopy data

Description

Train random forest model on hyperspec object

Usage

hs_RF(
  hs.x,
  metadata,
  target_var,
  spectrumID_col,
  ntree = 500,
  p_train = 0.75,
  metric = "Accuracy",
  ...
)

Arguments

hs.x	Hyperspec object
metadata	Dataframe containing the categorical variable/group to predict (target_var) and also a column with the matching spectrum identifiers (spectrumID_col).
target_var	Categorical variable/group to predict in metadata
spectrumID_col	Column with the matching spectrum identifiers (spectrumID_col) in metadata
ntree	Number of trees to build. Defaults to 500.
p_train	Percentage of data to use in training model. Defaults to 0.75.
metric	Metric to use to report/maximize performance of model (only for method_ML = "rf")
...	additional parameters passed on to caret::train

Examples

# Short example
data("hs_example")

# Preprocess
hs_example <- hs_preprocess(hs_example)

# Mock-up metadata
mock_meta <- data.frame(Spectrum_ID = rownames(hs_example@data$spc),
group = factor(c(rep(1,30),rep(2,34))))

# Calculate metrics
hs.RF <- hs_RF(hs.x = hs_example, metadata = mock_meta, spectrumID_col= "Spectrum_ID",
 target_var = "group")

 # Trained model
 print(hs.RF[[1]])

 # Confusion matrix
 print(hs.RF[[2]])

 # Variable importance metric
 caret::varImp(hs.RF[[1]])

 # Perform predictions
 hs_RF_pred(hs.x = hs_example, model = hs.RF[[1]])

CMET-UGent/MicroRaman documentation built on July 25, 2020, 6:20 p.m.