Package 'garchf'

Model-agnostic statistical probabilistic forecasting with conditional volatility

Description

Combines a mean model and a volatility model to produce probabilistic forecasts. Uncertainty is propagated from two sources: simulated innovation draws and simulated volatility paths derived from the variance model's residuals via bootstrap.

Usage

condvolf(
  y,
  h = 10L,
  mean_model = forecast::thetaf,
  sigma_model = forecast::thetaf,
  innovation = c("empirical", "gaussian", "student"),
  level = 95,
  B = 1000L,
  bootstrap_vol = TRUE,
  conformal = FALSE,
  cal_frac = 0.5,
  residuals_model = forecast::thetaf,
  ...
)

Arguments

y	A numeric vector or time series of class ts.
h	Forecasting horizon.
mean_model	Function to fit the mean model (default: forecast::auto.arima). Options include forecast::auto.arima, forecast::ets, forecast::thetaf, or any custom function that returns an object with forecast method.
sigma_model	Function to fit the variance model on squared residuals (default: forecast::auto.arima).
innovation	Distribution for standardized innovations. One of "gaussian" (fastest, often sufficient), "student" (heavy-tailed), or "empirical" (non-parametric, most flexible). Only used for conformal=FALSE
level	Confidence level for prediction intervals (default: 95).
B	Number of simulated paths (default: 2000).
bootstrap_vol	Whether to bootstrap volatility residuals (default: TRUE). If FALSE, uses parametric normal approximation for volatility.
conformal	Logical. If TRUE, enables split conformal prediction to calibrate interval width. The series is split into a training portion and a calibration portion (controlled by cal_frac); a base condvolf fit on the training portion produces calibration residuals, which are normalized by a volatility model, resampled via moving block bootstrap, re-inflated by the forecast volatility, and added to the full-data point forecast. This typically yields tighter and better- calibrated intervals than the raw simulation approach. Default is FALSE.
cal_frac	Numeric in (0, 1). Fraction of observations reserved for the calibration set when conformal = TRUE. The calibration set is always at least h observations long regardless of cal_frac. Larger values improve calibration stability at the cost of a shorter training set for the base model. Default is 0.5.
residuals_model	A function from package forecast, the model adjusted to residuals for 2-stage forecasting, and only for conformal=TRUE
...	Additional arguments passed to mean_model and sigma_model.

Value

A forecast object with components:

x

Original time series.

mean

Point forecast (mean of simulations).

lower

Lower prediction interval bound.

upper

Upper prediction interval bound.

level

Confidence level.

sims

Matrix of simulated forecast paths (h x B).

model

List containing mean and sigma model fits.

residuals

Residuals from the mean model.

standardized_residuals

Scaled standardized residuals used for innovation fitting.

method

Character string describing the method used.

Examples

library(forecast)

# Basic usage with Google stock data
y <- fpp2::goog200

# Gaussian innovations (fastest, often sufficient)
fc1 <- condvolf(y, h = 20, innovation = "gaussian")
plot(fc1)

# Student-t innovations (heavy-tailed)
fc2 <- condvolf(y, h = 20, innovation = "student")
plot(fc2)

# Empirical innovations (non-parametric)
fc3 <- condvolf(y, h = 20, innovation = "empirical")
plot(fc3)

# Using different mean and volatility models
fc4 <- condvolf(y, h = 20,
                mean_model = forecast::thetaf,
                sigma_model = forecast::ets,
                innovation = "gaussian")

# Compare prediction intervals
par(mfrow = c(2, 2))
plot(fc1, main = "Gaussian innovations")
plot(fc2, main = "Student-t innovations")
plot(fc3, main = "Empirical innovations")
plot(fc4, main = "Different mean and volatility models")
par(mfrow = c(1, 1))

---

GARCH forecasting wrapper returning a forecast object

Description

Similar spirit to forecast::thetaf(), but using rugarch models.

Usage

xgarchf(
  y,
  h = 10L,
  model = c("eGARCH", "sGARCH", "gjrGARCH", "apARCH", "iGARCH"),
  armaOrder = c(1, 1),
  garchOrder = c(1, 1),
  distribution.model = c("norm", "std", "ged"),
  level = c(80, 95),
  ...
)

Arguments

y	numeric vector or ts
h	forecast horizon
model	mean model ("sGARCH", "eGARCH", "gjrGARCH", ...)
armaOrder	ARMA order for the conditional mean
garchOrder	GARCH order
distribution.model	conditional distribution
level	confidence level(s)
...	Additional parameters passed to rugarch::ugarchfit

Value

object of class forecast

Examples

## Not run: 
y <- fpp2::goog200
fit <- xgarchf(y, h = 20, model = "eGARCH")
plot(fit)

## End(Not run)

---