Package edu.cnu.mdi.splot.fit

Class LinearLeastSquaresGuesser

java.lang.Object

edu.cnu.mdi.splot.fit.LinearLeastSquaresGuesser

public final class LinearLeastSquaresGuesser
extends Object

Utility for computing a weighted linear least-squares parameter guess: min sum_i w_i (A_i p - y_i)^2 by solving (sqrt(W)A)p = sqrt(W)y via QR.

Method Summary

Modifier and Type	Method	Description
static double[]	solve(double[][] A, double[] y, double[] weights)	Solve for parameters p in a linear model y ≈ A p using (optionally) weights.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Details

solve

public static double[] solve(double[][] A,
 double[] y,
 double[] weights)

Solve for parameters p in a linear model y ≈ A p using (optionally) weights. Rows are scaled by sqrt(weight_i) for WLS.

Parameters:

A - design matrix (n x p)

y - target vector (n)

weights - optional weights (n) or null for OLS

Returns:

parameter vector p (length p). Returns zeros if solve fails.