ARIMA

Rust包用于ARIMA模型系数估计和模拟。

示例

extern crate rand;
use rand::prelude::*;
use rand_distr::{Distribution, Normal};
use arima::{estimate, sim};

fn main() {
    // initialize RNG with seed
    let mut rng: StdRng = SeedableRng::from_seed([100; 32]);

    // our noise should be normally distributed
    let normal = Normal::new(10.0, 2.0).unwrap();

    // simulate time series
    let ts = sim::arima_sim(
        1000,                   // number of samples
        Some(&[0.7, 0.2]),      // AR parameters
        Some(&[0.4]),                   // MA parameters
        0,                      // difference parameter
        &|mut rng| { normal.sample(&mut rng) }, // noise fn
        &mut rng                // RNG
    ).unwrap();

    // estimate AR parameters
    let coef = estimate::fit(&ts, 2, 0, 1).unwrap();

    println!("Estimated parameters: {:?}", coef);
    // Estimated parameters: [14.904840907703845, 0.7524268545022731, 0.14075584488434256, 0.35966423499627603]
}

功能

• 完整的ARIMA模型参数估计
• 自相关/协方差计算
• 偏自相关计算
• AR参数估计
• 方差估计
• ARIMA时间序列模拟

路线图

• 订单估计

许可证

此包采用Apache-2.0许可证。