SensTimePlot function

Sensitivity analysis plot over time of the data

Plot of sensitivity of the neural network output respect to the inputs over the time variable from the data provided

SensTimePlot(
  object,
  fdata = NULL,
  date.var = NULL,
  facet = FALSE,
  smooth = FALSE,
  nspline = NULL,
  ...
)

Arguments

• object: fitted neural network model or array containing the raw sensitivities from the function SensAnalysisMLP

• fdata: data.frame containing the data to evaluate the sensitivity of the model. Not needed if the raw sensitivities has been passed as object

• date.var: Posixct vector with the date of each sample of fdata

  If NULL, the first variable with Posixct format of fdata is used as dates

• facet: logical if TRUE, function facet_grid from ggplot2 is used

• smooth: logical if TRUE, geom_smooth plots are added to each variable plot

• nspline: integer if smooth is TRUE, this determine the degree of the spline used to perform geom_smooth. If nspline is NULL, the square root of the length of the timeseries is used as degrees of the spline.

• ...: further arguments that should be passed to SensAnalysisMLP function

Returns

list of geom_line plots for the inputs variables representing the sensitivity of each output respect to the inputs over time

Examples

## Load data -------------------------------------------------------------------
data("DAILY_DEMAND_TR")
fdata <- DAILY_DEMAND_TR
fdata[,3] <- ifelse(as.data.frame(fdata)[,3] %in% c("SUN","SAT"), 0, 1)
## Parameters of the NNET ------------------------------------------------------
hidden_neurons <- 5
iters <- 250
decay <- 0.1

################################################################################
#########################  REGRESSION NNET #####################################
################################################################################
## Regression dataframe --------------------------------------------------------
# Scale the data
fdata.Reg.tr <- fdata[,2:ncol(fdata)]
fdata.Reg.tr[,3] <- fdata.Reg.tr[,3]/10
fdata.Reg.tr[,1] <- fdata.Reg.tr[,1]/1000

# Normalize the data for some models
preProc <- caret::preProcess(fdata.Reg.tr, method = c("center","scale"))
nntrData <- predict(preProc, fdata.Reg.tr)

#' ## TRAIN nnet NNET --------------------------------------------------------
# Create a formula to train NNET
form <- paste(names(fdata.Reg.tr)[2:ncol(fdata.Reg.tr)], collapse = " + ")
form <- formula(paste(names(fdata.Reg.tr)[1], form, sep = " ~ "))

set.seed(150)
nnetmod <- nnet::nnet(form,
                      data = nntrData,
                      linear.output = TRUE,
                      size = hidden_neurons,
                      decay = decay,
                      maxit = iters)
# Try SensTimePlot
NeuralSens::SensTimePlot(nnetmod, fdata = nntrData, date.var = NULL)

References

Pizarroso J, Portela J, Muñoz A (2022). NeuralSens: Sensitivity Analysis of Neural Networks. Journal of Statistical Software, 102(7), 1-36.