Struct FGLS 

pub struct FGLS;

Implementations

impl FGLS

pub fn wls_from_formula( formula: &Formula, data: &DataFrame, weights: &Array1<f64>, ) -> Result<FglsResult, GreenersError>

Weighted Least Squares (WLS) from a formula and DataFrame.

Examples

use greeners::{FGLS, DataFrame, Formula};
use ndarray::Array1;
use std::collections::HashMap;

let mut data = HashMap::new();
data.insert("y".to_string(), Array1::from(vec![1.0, 2.0, 3.0]));
data.insert("x1".to_string(), Array1::from(vec![1.0, 2.0, 3.0]));
data.insert("weight".to_string(), Array1::from(vec![1.0, 1.0, 1.0]));

let df = DataFrame::new(data).unwrap();
let formula = Formula::parse("y ~ x1").unwrap();
let weights = df.get("weight").unwrap();

let result = FGLS::wls_from_formula(&formula, &df, weights).unwrap();

pub fn wls( y: &Array1<f64>, x: &Array2<f64>, weights: &Array1<f64>, ) -> Result<FglsResult, GreenersError>

Weighted Least Squares (WLS) Used when there is known or estimated heteroscedasticity. Weights (w) should be inversely proportional to variance: w_i = 1 / sigma_i^2

pub fn wls_with_names( y: &Array1<f64>, x: &Array2<f64>, weights: &Array1<f64>, variable_names: Option<Vec<String>>, ) -> Result<FglsResult, GreenersError>

pub fn cochrane_orcutt_from_formula( formula: &Formula, data: &DataFrame, ) -> Result<FglsResult, GreenersError>

Cochrane-Orcutt Iterative Procedure (AR(1)) from a formula and DataFrame.

pub fn cochrane_orcutt( y: &Array1<f64>, x: &Array2<f64>, ) -> Result<FglsResult, GreenersError>

Cochrane-Orcutt Iterative Procedure (AR(1)) Solves serial correlation: u_t = rho * u_{t-1} + e_t Recovers the efficiency (BLUE) that OLS loses.

pub fn cochrane_orcutt_with_names( y: &Array1<f64>, x: &Array2<f64>, variable_names: Option<Vec<String>>, ) -> Result<FglsResult, GreenersError>