[284] | 1 | <?php
|
---|
[2] | 2 | /*=======================================================================
|
---|
[284] | 3 | // File: JPGRAPH_REGSTAT.PHP
|
---|
| 4 | // Description: Regression and statistical analysis helper classes
|
---|
| 5 | // Created: 2002-12-01
|
---|
| 6 | // Ver: $Id: jpgraph_regstat.php 1131 2009-03-11 20:08:24Z ljp $
|
---|
| 7 | //
|
---|
| 8 | // Copyright (c) Asial Corporation. All rights reserved.
|
---|
| 9 | //========================================================================
|
---|
| 10 | */
|
---|
[2] | 11 |
|
---|
| 12 | //------------------------------------------------------------------------
|
---|
| 13 | // CLASS Spline
|
---|
| 14 | // Create a new data array from an existing data array but with more points.
|
---|
| 15 | // The new points are interpolated using a cubic spline algorithm
|
---|
| 16 | //------------------------------------------------------------------------
|
---|
| 17 | class Spline {
|
---|
| 18 | // 3:rd degree polynom approximation
|
---|
| 19 |
|
---|
| 20 | private $xdata,$ydata; // Data vectors
|
---|
[284] | 21 | private $y2; // 2:nd derivate of ydata
|
---|
[2] | 22 | private $n=0;
|
---|
| 23 |
|
---|
[284] | 24 | function __construct($xdata,$ydata) {
|
---|
| 25 | $this->y2 = array();
|
---|
| 26 | $this->xdata = $xdata;
|
---|
| 27 | $this->ydata = $ydata;
|
---|
[2] | 28 |
|
---|
[284] | 29 | $n = count($ydata);
|
---|
| 30 | $this->n = $n;
|
---|
| 31 | if( $this->n !== count($xdata) ) {
|
---|
| 32 | JpGraphError::RaiseL(19001);
|
---|
| 33 | //('Spline: Number of X and Y coordinates must be the same');
|
---|
| 34 | }
|
---|
[2] | 35 |
|
---|
[284] | 36 | // Natural spline 2:derivate == 0 at endpoints
|
---|
| 37 | $this->y2[0] = 0.0;
|
---|
| 38 | $this->y2[$n-1] = 0.0;
|
---|
| 39 | $delta[0] = 0.0;
|
---|
[2] | 40 |
|
---|
[284] | 41 | // Calculate 2:nd derivate
|
---|
| 42 | for($i=1; $i < $n-1; ++$i) {
|
---|
| 43 | $d = ($xdata[$i+1]-$xdata[$i-1]);
|
---|
| 44 | if( $d == 0 ) {
|
---|
| 45 | JpGraphError::RaiseL(19002);
|
---|
| 46 | //('Invalid input data for spline. Two or more consecutive input X-values are equal. Each input X-value must differ since from a mathematical point of view it must be a one-to-one mapping, i.e. each X-value must correspond to exactly one Y-value.');
|
---|
| 47 | }
|
---|
| 48 | $s = ($xdata[$i]-$xdata[$i-1])/$d;
|
---|
| 49 | $p = $s*$this->y2[$i-1]+2.0;
|
---|
| 50 | $this->y2[$i] = ($s-1.0)/$p;
|
---|
| 51 | $delta[$i] = ($ydata[$i+1]-$ydata[$i])/($xdata[$i+1]-$xdata[$i]) -
|
---|
| 52 | ($ydata[$i]-$ydata[$i-1])/($xdata[$i]-$xdata[$i-1]);
|
---|
| 53 | $delta[$i] = (6.0*$delta[$i]/($xdata[$i+1]-$xdata[$i-1])-$s*$delta[$i-1])/$p;
|
---|
| 54 | }
|
---|
[2] | 55 |
|
---|
[284] | 56 | // Backward substitution
|
---|
| 57 | for( $j=$n-2; $j >= 0; --$j ) {
|
---|
| 58 | $this->y2[$j] = $this->y2[$j]*$this->y2[$j+1] + $delta[$j];
|
---|
| 59 | }
|
---|
[2] | 60 | }
|
---|
| 61 |
|
---|
| 62 | // Return the two new data vectors
|
---|
| 63 | function Get($num=50) {
|
---|
[284] | 64 | $n = $this->n ;
|
---|
| 65 | $step = ($this->xdata[$n-1]-$this->xdata[0]) / ($num-1);
|
---|
| 66 | $xnew=array();
|
---|
| 67 | $ynew=array();
|
---|
| 68 | $xnew[0] = $this->xdata[0];
|
---|
| 69 | $ynew[0] = $this->ydata[0];
|
---|
| 70 | for( $j=1; $j < $num; ++$j ) {
|
---|
| 71 | $xnew[$j] = $xnew[0]+$j*$step;
|
---|
| 72 | $ynew[$j] = $this->Interpolate($xnew[$j]);
|
---|
| 73 | }
|
---|
| 74 | return array($xnew,$ynew);
|
---|
[2] | 75 | }
|
---|
| 76 |
|
---|
| 77 | // Return a single interpolated Y-value from an x value
|
---|
| 78 | function Interpolate($xpoint) {
|
---|
| 79 |
|
---|
[284] | 80 | $max = $this->n-1;
|
---|
| 81 | $min = 0;
|
---|
[2] | 82 |
|
---|
[284] | 83 | // Binary search to find interval
|
---|
| 84 | while( $max-$min > 1 ) {
|
---|
| 85 | $k = ($max+$min) / 2;
|
---|
| 86 | if( $this->xdata[$k] > $xpoint )
|
---|
| 87 | $max=$k;
|
---|
| 88 | else
|
---|
| 89 | $min=$k;
|
---|
| 90 | }
|
---|
[2] | 91 |
|
---|
[284] | 92 | // Each interval is interpolated by a 3:degree polynom function
|
---|
| 93 | $h = $this->xdata[$max]-$this->xdata[$min];
|
---|
[2] | 94 |
|
---|
[284] | 95 | if( $h == 0 ) {
|
---|
| 96 | JpGraphError::RaiseL(19002);
|
---|
| 97 | //('Invalid input data for spline. Two or more consecutive input X-values are equal. Each input X-value must differ since from a mathematical point of view it must be a one-to-one mapping, i.e. each X-value must correspond to exactly one Y-value.');
|
---|
| 98 | }
|
---|
[2] | 99 |
|
---|
| 100 |
|
---|
[284] | 101 | $a = ($this->xdata[$max]-$xpoint)/$h;
|
---|
| 102 | $b = ($xpoint-$this->xdata[$min])/$h;
|
---|
| 103 | return $a*$this->ydata[$min]+$b*$this->ydata[$max]+
|
---|
| 104 | (($a*$a*$a-$a)*$this->y2[$min]+($b*$b*$b-$b)*$this->y2[$max])*($h*$h)/6.0;
|
---|
[2] | 105 | }
|
---|
| 106 | }
|
---|
| 107 |
|
---|
| 108 | //------------------------------------------------------------------------
|
---|
| 109 | // CLASS Bezier
|
---|
| 110 | // Create a new data array from a number of control points
|
---|
| 111 | //------------------------------------------------------------------------
|
---|
| 112 | class Bezier {
|
---|
[284] | 113 | /**
|
---|
| 114 | * @author Thomas Despoix, openXtrem company
|
---|
| 115 | * @license released under QPL
|
---|
| 116 | * @abstract Bezier interoplated point generation,
|
---|
| 117 | * computed from control points data sets, based on Paul Bourke algorithm :
|
---|
| 118 | * http://local.wasp.uwa.edu.au/~pbourke/geometry/bezier/index2.html
|
---|
| 119 | */
|
---|
[2] | 120 | private $datax = array();
|
---|
| 121 | private $datay = array();
|
---|
| 122 | private $n=0;
|
---|
[284] | 123 |
|
---|
| 124 | function __construct($datax, $datay, $attraction_factor = 1) {
|
---|
| 125 | // Adding control point multiple time will raise their attraction power over the curve
|
---|
| 126 | $this->n = count($datax);
|
---|
| 127 | if( $this->n !== count($datay) ) {
|
---|
| 128 | JpGraphError::RaiseL(19003);
|
---|
| 129 | //('Bezier: Number of X and Y coordinates must be the same');
|
---|
| 130 | }
|
---|
| 131 | $idx=0;
|
---|
| 132 | foreach($datax as $datumx) {
|
---|
| 133 | for ($i = 0; $i < $attraction_factor; $i++) {
|
---|
| 134 | $this->datax[$idx++] = $datumx;
|
---|
| 135 | }
|
---|
| 136 | }
|
---|
| 137 | $idx=0;
|
---|
| 138 | foreach($datay as $datumy) {
|
---|
| 139 | for ($i = 0; $i < $attraction_factor; $i++) {
|
---|
| 140 | $this->datay[$idx++] = $datumy;
|
---|
| 141 | }
|
---|
| 142 | }
|
---|
| 143 | $this->n *= $attraction_factor;
|
---|
[2] | 144 | }
|
---|
| 145 |
|
---|
[284] | 146 | /**
|
---|
| 147 | * Return a set of data points that specifies the bezier curve with $steps points
|
---|
| 148 | * @param $steps Number of new points to return
|
---|
| 149 | * @return array($datax, $datay)
|
---|
| 150 | */
|
---|
[2] | 151 | function Get($steps) {
|
---|
[284] | 152 | $datax = array();
|
---|
| 153 | $datay = array();
|
---|
| 154 | for ($i = 0; $i < $steps; $i++) {
|
---|
| 155 | list($datumx, $datumy) = $this->GetPoint((double) $i / (double) $steps);
|
---|
| 156 | $datax[$i] = $datumx;
|
---|
| 157 | $datay[$i] = $datumy;
|
---|
| 158 | }
|
---|
| 159 |
|
---|
| 160 | $datax[] = end($this->datax);
|
---|
| 161 | $datay[] = end($this->datay);
|
---|
| 162 |
|
---|
| 163 | return array($datax, $datay);
|
---|
[2] | 164 | }
|
---|
[284] | 165 |
|
---|
| 166 | /**
|
---|
| 167 | * Return one point on the bezier curve. $mu is the position on the curve where $mu is in the
|
---|
| 168 | * range 0 $mu < 1 where 0 is tha start point and 1 is the end point. Note that every newly computed
|
---|
| 169 | * point depends on all the existing points
|
---|
| 170 | *
|
---|
| 171 | * @param $mu Position on the bezier curve
|
---|
| 172 | * @return array($x, $y)
|
---|
| 173 | */
|
---|
[2] | 174 | function GetPoint($mu) {
|
---|
[284] | 175 | $n = $this->n - 1;
|
---|
| 176 | $k = 0;
|
---|
| 177 | $kn = 0;
|
---|
| 178 | $nn = 0;
|
---|
| 179 | $nkn = 0;
|
---|
| 180 | $blend = 0.0;
|
---|
| 181 | $newx = 0.0;
|
---|
| 182 | $newy = 0.0;
|
---|
[2] | 183 |
|
---|
[284] | 184 | $muk = 1.0;
|
---|
| 185 | $munk = (double) pow(1-$mu,(double) $n);
|
---|
[2] | 186 |
|
---|
[284] | 187 | for ($k = 0; $k <= $n; $k++) {
|
---|
| 188 | $nn = $n;
|
---|
| 189 | $kn = $k;
|
---|
| 190 | $nkn = $n - $k;
|
---|
| 191 | $blend = $muk * $munk;
|
---|
| 192 | $muk *= $mu;
|
---|
| 193 | $munk /= (1-$mu);
|
---|
| 194 | while ($nn >= 1) {
|
---|
| 195 | $blend *= $nn;
|
---|
| 196 | $nn--;
|
---|
| 197 | if ($kn > 1) {
|
---|
| 198 | $blend /= (double) $kn;
|
---|
| 199 | $kn--;
|
---|
| 200 | }
|
---|
| 201 | if ($nkn > 1) {
|
---|
| 202 | $blend /= (double) $nkn;
|
---|
| 203 | $nkn--;
|
---|
| 204 | }
|
---|
| 205 | }
|
---|
| 206 | $newx += $this->datax[$k] * $blend;
|
---|
| 207 | $newy += $this->datay[$k] * $blend;
|
---|
| 208 | }
|
---|
[2] | 209 |
|
---|
[284] | 210 | return array($newx, $newy);
|
---|
[2] | 211 | }
|
---|
| 212 | }
|
---|
| 213 |
|
---|
| 214 | // EOF
|
---|
| 215 | ?>
|
---|