semiconditional_train_valid_GPD_loss: Semi-conditional GPD MLEs and their train-validation...
In EQRN: Extreme Quantile Regression Neural Networks for Risk Forecasting

semiconditional_train_valid_GPD_loss

R Documentation

Semi-conditional GPD MLEs and their train-validation likelihoods

Semi-conditional GPD MLEs and their train-validation likelihoods

semiconditional_train_valid_GPD_loss(
  Y_train,
  Y_valid,
  interm_quant_train,
  interm_quant_valid
)

`Y_train`	Vector of "training" observations on which to estimate the MLEs.
`Y_valid`	Vector of "validation" observations, on which to estimate the out of training sample GPD loss.
`interm_quant_train`	Vector of intermediate quantiles serving as a varying threshold for each training observation.
`interm_quant_valid`	Vector of intermediate quantiles serving as a varying threshold for each validation observation.

Named list containing:

`scale`	GPD scale MLE inferred from the train set,
`shape`	GPD shape MLE inferred from the train set,
`train_loss`	the negative log-likelihoods of the MLEs over the training samples,
`valid_loss`	the negative log-likelihoods of the MLEs over the validation samples.