Struct ARIMA 

pub struct ARIMA { /* private fields */ }

ARIMA (Auto-Regressive Integrated Moving Average) model.

ARIMA(p, d, q) model for time series forecasting:

• p: Order of auto-regressive (AR) component
• d: Degree of differencing (I) for stationarity
• q: Order of moving average (MA) component

Examples

use aprender::time_series::ARIMA;
use aprender::primitives::Vector;

// Create ARIMA(1, 1, 1) model
let mut model = ARIMA::new(1, 1, 1);

// Fit to time series data
let data = Vector::from_slice(&[10.0, 12.0, 13.0, 15.0, 14.0, 16.0, 18.0]);
model.fit(&data).expect("fit should succeed");

// Forecast next period
let forecast = model.forecast(1).expect("forecast should succeed");
assert!(forecast.len() == 1);

Implementations

impl ARIMA

pub fn new(p: usize, d: usize, q: usize) -> ARIMA

Create a new ARIMA model with specified orders.

Arguments

• p - AR order (number of lagged observations)
• d - Differencing order (degree of differencing)
• q - MA order (number of lagged forecast errors)

Examples

use aprender::time_series::ARIMA;

// ARIMA(1, 1, 1) - simple model
let model = ARIMA::new(1, 1, 1);

// ARIMA(2, 1, 0) - AR(2) with differencing
let ar_model = ARIMA::new(2, 1, 0);

// ARIMA(0, 1, 1) - MA(1) with differencing (equivalent to exponential smoothing)
let ma_model = ARIMA::new(0, 1, 1);

pub fn fit(&mut self, data: &Vector<f64>) -> Result<(), AprenderError>

Fit the ARIMA model to time series data.

This performs differencing (if d > 0) and estimates AR/MA parameters using least squares estimation.

Arguments

• data - Time series observations

Returns

• Ok(()) - Model fitted successfully
• Err(AprenderError) - If fitting fails (insufficient data, etc.)

Examples

use aprender::time_series::ARIMA;
use aprender::primitives::Vector;

let mut model = ARIMA::new(1, 1, 1);
let data = Vector::from_slice(&[10.0, 12.0, 13.0, 15.0, 14.0, 16.0]);
model.fit(&data).expect("fit should succeed");

pub fn forecast(&self, n_periods: usize) -> Result<Vector<f64>, AprenderError>

Forecast future values.

Generates forecasts for the next n_periods using the fitted model.

Arguments

• n_periods - Number of periods to forecast

Returns

• Ok(Vector) - Forecasted values
• Err(AprenderError) - If model hasn’t been fitted or forecasting fails

Examples

use aprender::time_series::ARIMA;
use aprender::primitives::Vector;

let mut model = ARIMA::new(1, 1, 1);
let data = Vector::from_slice(&[10.0, 12.0, 13.0, 15.0, 14.0, 16.0]);
model.fit(&data).expect("fit should succeed");

// Forecast next 3 periods
let forecast = model.forecast(3).expect("forecast should succeed");
assert_eq!(forecast.len(), 3);

pub fn ar_coefficients(&self) -> Option<&Vector<f64>>

Get the AR coefficients (if fitted).

Examples

use aprender::time_series::ARIMA;
use aprender::primitives::Vector;

let mut model = ARIMA::new(2, 0, 0);
let data = Vector::from_slice(&[1.0, 2.0, 3.0, 4.0, 5.0, 6.0]);
model.fit(&data).expect("fit should succeed");

let ar_coef = model.ar_coefficients().expect("should have AR coefficients");
assert_eq!(ar_coef.len(), 2);

pub fn ma_coefficients(&self) -> Option<&Vector<f64>>

Get the MA coefficients (if fitted).

Examples

use aprender::time_series::ARIMA;
use aprender::primitives::Vector;

let mut model = ARIMA::new(0, 0, 1);
let data = Vector::from_slice(&[1.0, 2.0, 3.0, 4.0, 5.0, 6.0]);
model.fit(&data).expect("fit should succeed");

let ma_coef = model.ma_coefficients().expect("should have MA coefficients");
assert_eq!(ma_coef.len(), 1);

pub fn intercept(&self) -> f64

Get the model intercept (if fitted).

pub fn order(&self) -> (usize, usize, usize)

Get the model order (p, d, q).

Examples

use aprender::time_series::ARIMA;

let model = ARIMA::new(1, 1, 1);
assert_eq!(model.order(), (1, 1, 1));

Trait Implementations

impl Clone for ARIMA

fn clone(&self) -> ARIMA

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ARIMA

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.