Extreme Quantile Regression Neural Networks for Risk Forecasting

---

Documentation for package ‘EQRN’ version 0.1.1

• DESCRIPTION file.
• Package NEWS.

Help Pages

check_directory	Check directory existence
compute_EQRN_GPDLoss	Generalized Pareto likelihood loss of a EQRN_iid predictor
compute_EQRN_seq_GPDLoss	Generalized Pareto likelihood loss of a EQRN_seq predictor
default_device	Default torch device
end_doFuture_strategy	End the currently set doFuture strategy
EQRN_excess_probability	Tail excess probability prediction using an EQRN_iid object
EQRN_excess_probability_seq	Tail excess probability prediction using an EQRN_seq object
EQRN_fit	EQRN fit function for independent data
EQRN_fit_restart	Wrapper for fitting EQRN with restart for stability
EQRN_fit_seq	EQRN fit function for sequential and time series data
EQRN_load	Load an EQRN object from disc
EQRN_predict	Predict function for an EQRN_iid fitted object
EQRN_predict_params	GPD parameters prediction function for an EQRN_iid fitted object
EQRN_predict_params_seq	GPD parameters prediction function for an EQRN_seq fitted object
EQRN_predict_seq	Predict function for an EQRN_seq fitted object
EQRN_save	Save an EQRN object on disc
excess_probability	Excess Probability Predictions
excess_probability.EQRN_iid	Tail excess probability prediction method using an EQRN_iid object
excess_probability.EQRN_seq	Tail excess probability prediction method using an EQRN_iid object
FC_GPD_net	MLP module for GPD parameter prediction
FC_GPD_SNN	Self-normalized fully-connected network module for GPD parameter prediction
fit_GPD_unconditional	Maximum likelihood estimates for the GPD distribution using peaks over threshold
get_doFuture_operator	Get doFuture operator
get_excesses	Computes rescaled excesses over the conditional quantiles
GPD_excess_probability	Tail excess probability prediction based on conditional GPD parameters
GPD_quantiles	Compute extreme quantile from GPD parameters
install_backend	Install Torch Backend
lagged_features	Covariate lagged replication for temporal dependence
last_elem	Last element of a vector
loss_GPD	Generalized Pareto likelihood loss
loss_GPD_tensor	GPD tensor loss function for training a EQRN network
make_folds	Create cross-validation folds
mean_absolute_error	Mean absolute error
mean_squared_error	Mean squared error
mts_dataset	Dataset creator for sequential data
multilevel_exceedance_proba_error	Multilevel 'quantile_exceedance_proba_error'
multilevel_MAE	Multilevel quantile MAEs
multilevel_MSE	Multilevel quantile MSEs
multilevel_pred_bias	Multilevel prediction bias
multilevel_prop_below	Multilevel 'proportion_below'
multilevel_q_loss	Multilevel quantile losses
multilevel_q_pred_error	Multilevel 'quantile_prediction_error'
multilevel_resid_var	Multilevel residual variance
multilevel_R_squared	Multilevel R squared
perform_scaling	Performs feature scaling without overfitting
predict.EQRN_iid	Predict method for an EQRN_iid fitted object
predict.EQRN_seq	Predict method for an EQRN_seq fitted object
predict.QRN_seq	Predict method for a QRN_seq fitted object
prediction_bias	Prediction bias
prediction_residual_variance	Prediction residual variance
predict_GPD_semiconditional	Predict semi-conditional extreme quantiles using peaks over threshold
predict_unconditional_quantiles	Predict unconditional extreme quantiles using peaks over threshold
process_features	Feature processor for EQRN
proportion_below	Proportion of observations below conditional quantile vector
QRNN_RNN_net	Recurrent quantile regression neural network module
QRN_fit_multiple	Wrapper for fitting a recurrent QRN with restart for stability
QRN_seq_fit	Recurrent QRN fitting function
QRN_seq_predict	Predict function for a QRN_seq fitted object
QRN_seq_predict_foldwise	Foldwise fit-predict function using a recurrent QRN
QRN_seq_predict_foldwise_sep	Sigle-fold foldwise fit-predict function using a recurrent QRN
quantile_exceedance_proba_error	Quantile exceedance probability prediction calibration error
quantile_loss	Quantile loss
quantile_loss_tensor	Tensor quantile loss function for training a QRN network
quantile_prediction_error	Quantile prediction calibration error
Recurrent_GPD_net	Recurrent network module for GPD parameter prediction
roundm	Mathematical number rounding
R_squared	R squared
safe_save_rds	Safe RDS save
semiconditional_train_valid_GPD_loss	Semi-conditional GPD MLEs and their train-validation likelihoods
Separated_GPD_SNN	Self-normalized separated network module for GPD parameter prediction
set_doFuture_strategy	Set a doFuture execution strategy
square_loss	Square loss
unconditional_train_valid_GPD_loss	Unconditional GPD MLEs and their train-validation likelihoods
vec2mat	Convert a vector to a matrix
vector_insert	Insert value in vector