Unit root tests in Rust

Stationarity tests for time-series data. It's just Dickey-Fuller yet, but it's a start.

Usage

```rust use unitroot::prelude::distrib::dickeyfuller::model1approxcriticalvalue; use unitroot::prelude::distrib::AlphaLevel; use unitroot::prelude::nalgebra::DVector; use unitroot::prelude::*;

fn main() { let y = DVector::fromrowslice(&[ -0.89642362, 0.3222552, -1.96581989, -1.10012936, -1.3682928, 1.17239875, 2.19561259, 2.54295031, 2.05530587, 1.13212955, -0.42968979, ]);

// compute the test statistic
let report = tools::univariate_dickeyfuller(&y);

// critical values for model 1 (constant but no trend):
let critical_value = model_1_approx_critical_value(report.size, AlphaLevel::OnePercent);
assert_eq!(report.size, 10);

// comparison
let t_stat = report.test_statistic.unwrap();
println!("t-statistic: {}", t_stat);
assert!((t_stat - -1.472691).abs() < 1e-6);
assert!(t_stat > critical_value);

} ```