customscale: Scale training and test datasets for anomaly detection

In SpencerButt/IDPS-LAAD: Intrusion Detection Prevention System Log Autoencoder Anomaly Detector

Description

Scale training and test datasets for anomaly detection

Usage

customscale(DF_train, DF_test, a, b)

Arguments

DF_train	The unscaled training dataset used to scale the values in the test dataset.
DF_test	The unscaled test dataset scaled to the user defined interval [a, b].
a	The minimum desired scaling value. This will be the minimum value of the scaled training dataset. The test dataset minimum may be smaller than this value.
b	The maximum desired scaling value. This will be the maximum value of the scaled training dataset. The test dataset maximum may be larger than this value.

Value

ScaleDF The DF_test dataset scaled to [a,b] as determined from the DF_train dataset

Warning

Requires that the test and training datasets have the same numeric features in the same order

Examples

## Not run: 
# Select only numeric features
irisExample <- iris[ ,1:4]

# Add column to split 80% training and 20% test
splitIris <- irisExample %>%
  dplyr::mutate(Cross = sample(c(1, 2),
                              size = nrow(irisExample),
                              replace = TRUE,
                              prob = c(0.80, 0.20)))

# Obtain training dataset
irisTrain <- splitIris[splitIris$Cross == 1, ] %>%                                                             
  dplyr::select(-Cross)


# Obtain test dataset
irisTest <- splitIris[splitIris$Cross == 2, ] %>%
  dplyr::select(-Cross)

# Scale irisTrain into [0, 1] (irisTrainZO)
irisTrainZO <- customscale(irisTrain, irisTrain, 0, 1)

# Scale irisTest into [0, 1] using 
irisTest <- customscale(irisTrain, irisTest, 0 ,1)

## End(Not run)

SpencerButt/IDPS-LAAD documentation built on April 20, 2020, 8:45 p.m.