Lapack::leastSquaresBySVD

(PECL lapack >= 0.1.0)

Lapack::leastSquaresBySVD — Solve the linear least squares problem, using SVD

説明

public static array Lapack::leastSquaresBySVD ( array $a , array $b )

Solve the linear least squares problem, find min x in || B - Ax || Returns an array representing x. Expects arrays of arrays, and will return an array of arrays in the dimension B num cols x A num cols. Uses SVD with a divide and conquer algorithm.

パラメータ

a: Matrix A
b: Matrix B

返り値

Returns the solution as an array of arrays.

例

例1 Using Lapack::leastSquaresBySVD():


<?php

  $a = array(
      array( 1.44,  -7.84,  -4.39,   4.53),
      array(-9.96,  -0.28,  -3.24,   3.83),
      array(-7.55,   3.24,   6.27,  -6.64),
      array( 8.34,   8.09,   5.28,   2.06),
      array( 7.08,   2.52,   0.74,  -2.47),
      array(-5.45,  -5.70,  -1.19,   4.70),
  );

  $b = array(
      array( 8.58,   9.35),
      array( 8.26,  -4.43),
      array( 8.48,  -0.70),
      array(-5.28,  -0.26),
      array( 5.72,  -7.36),
      array( 8.93,  -2.52),           
  );

  $result = Lapack::leastSquaresBySVD($a, $b);
  
  ?>