ARMA model

R6 class for ARMA(p, q) model

Note

Version 1.0.2

See also

stats::arima() which is wrapped in the method fit.

Public fields

control

list, to contain custom control parameters.

model_name

character(1), standard model name.

Active bindings

arOrder

integer(1), Autoregressive order.

maOrder

integer(1), Moving-Average order.

order

named numeric vector of size 2. Orders of the ARMA model: the first element is the arOrder, while the second the maOrder.

intercept

named integer(1), intercept of the model.

phi

named numeric vector of size arOrder, AR parameters. If arOrder = 0, the parameter is zero.

theta

named numeric vector of size maOrder, MA parameters. If maOrder = 0, the parameter is zero.

coefficients

named numeric vector of size arOrder + maOrder + 1. The first element ie sthe intercept, then the ARMA parameters.

std.errors

Numeric named vector, standard errors of the intercept and ARMA parameters.

sigma2

integer(1), std.errors of the residuals.

A

numeric matrix of size (arOrder + maOrder) x (arOrder + maOrder). See the function ARMA_companion_matrix for more details.

b

numeric vector of size arOrder + maOrder. See the function ARMA_vector_b for more details.

tidy

Tibble with estimated parameters and standard errors.

Methods

Public methods

• ARMA_modelR6$new()

• ARMA_modelR6$fit()

• ARMA_modelR6$filter()

• ARMA_modelR6$next_step()

• ARMA_modelR6$expectation()

• ARMA_modelR6$variance()

• ARMA_modelR6$update()

• ARMA_modelR6$update_std.errors()

• ARMA_modelR6$update_sigma2()

• ARMA_modelR6$print()

• ARMA_modelR6$clone()

---

Method new()

Initialize an ARMA model

Usage

ARMA_modelR6$new(arOrder = 1, maOrder = 1, include.intercept = FALSE)

Arguments

arOrder

integer(1), order of the Autoregressive component.

maOrder

integer(1), order of the Moving-Average component.

include.intercept

logical(1), the default is FALSE. When TRUE the intercept will be included.

---

Method fit()

Fit the ARMA model with arima function.

Usage

ARMA_modelR6$fit(x)

Arguments

x

numeric vector, time series to fit.

---

Method filter()

Filter the time-series and compute fitted values and residuals. See the function ARMA_filter for more details.

Usage

ARMA_modelR6$filter(x)

Arguments

x

numeric vector, time series to filter.

---

Method next_step()

Next step function. See the function ARMA_next_step for more details.

Usage

ARMA_modelR6$next_step(x, n.ahead = 1, eps = 0)

Arguments

x

Numeric vector, state vector with past observations and residuals.

n.ahead

integer(1), number of steps ahead.

eps

optional numeric vector of length n.ahead, next step realized residuals.

---

Method expectation()

Forecast expected value. See the function ARMA_expectation for more details.

Usage

ARMA_modelR6$expectation(h = 1, X0)

Arguments

h

integer(1), number of steps ahead.

X0

numeric vector of size p + q. State vector of past values.

---

Method variance()

Forecast variance. See the function ARMA_variance for more details.

Usage

ARMA_modelR6$variance(h = 1, sigma2 = 1)

Arguments

h

integer(1), number of steps ahead.

sigma2

integer(1), standard deviation of the residuals.

---

Method update()

Update the model's parameters

Usage

ARMA_modelR6$update(coefficients)

Arguments

coefficients

Numeric named vector, model's coefficients. If missing nothing will be updated.

---

Method update_std.errors()

Update the standard errors of the parameters.

Usage

ARMA_modelR6$update_std.errors(std.errors)

Arguments

std.errors

Numeric named vector, parameters' standard errors. If missing nothing will be updated.

---

Method update_sigma2()

Update the variance of the residuals.

Usage

ARMA_modelR6$update_sigma2(sigma2)

Arguments

sigma2

integer(1), variance of the residuals.

---

Method print()

Print method for ARMA_modelR6 class.

Usage

ARMA_modelR6$print()

---

Method clone()

The objects of this class are cloneable with this method.

Usage

ARMA_modelR6$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.