You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
wkcrm/vendor/phpexcel/PHPExcel/Shared/trend/bestFitClass.php

426 lines
12 KiB

<?php
/**
* PHPExcel_Best_Fit
*
* Copyright (c) 2006 - 2015 PHPExcel
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*
* @category PHPExcel
* @package PHPExcel_Shared_Trend
* @copyright Copyright (c) 2006 - 2015 PHPExcel (http://www.codeplex.com/PHPExcel)
* @license http://www.gnu.org/licenses/old-licenses/lgpl-2.1.txt LGPL
* @version ##VERSION##, ##DATE##
*/
class PHPExcel_Best_Fit
{
/**
* Indicator flag for a calculation error
*
* @var boolean
**/
protected $error = false;
/**
* Algorithm type to use for best-fit
*
* @var string
**/
protected $bestFitType = 'undetermined';
/**
* Number of entries in the sets of x- and y-value arrays
*
* @var int
**/
protected $valueCount = 0;
/**
* X-value dataseries of values
*
* @var float[]
**/
protected $xValues = array();
/**
* Y-value dataseries of values
*
* @var float[]
**/
protected $yValues = array();
/**
* Flag indicating whether values should be adjusted to Y=0
*
* @var boolean
**/
protected $adjustToZero = false;
/**
* Y-value series of best-fit values
*
* @var float[]
**/
protected $yBestFitValues = array();
protected $goodnessOfFit = 1;
protected $stdevOfResiduals = 0;
protected $covariance = 0;
protected $correlation = 0;
protected $SSRegression = 0;
protected $SSResiduals = 0;
protected $DFResiduals = 0;
protected $f = 0;
protected $slope = 0;
protected $slopeSE = 0;
protected $intersect = 0;
protected $intersectSE = 0;
protected $xOffset = 0;
protected $yOffset = 0;
public function getError()
{
return $this->error;
}
public function getBestFitType()
{
return $this->bestFitType;
}
/**
* Return the Y-Value for a specified value of X
*
* @param float $xValue X-Value
* @return float Y-Value
*/
public function getValueOfYForX($xValue)
{
return false;
}
/**
* Return the X-Value for a specified value of Y
*
* @param float $yValue Y-Value
* @return float X-Value
*/
public function getValueOfXForY($yValue)
{
return false;
}
/**
* Return the original set of X-Values
*
* @return float[] X-Values
*/
public function getXValues()
{
return $this->xValues;
}
/**
* Return the Equation of the best-fit line
*
* @param int $dp Number of places of decimal precision to display
* @return string
*/
public function getEquation($dp = 0)
{
return false;
}
/**
* Return the Slope of the line
*
* @param int $dp Number of places of decimal precision to display
* @return string
*/
public function getSlope($dp = 0)
{
if ($dp != 0) {
return round($this->slope, $dp);
}
return $this->slope;
}
/**
* Return the standard error of the Slope
*
* @param int $dp Number of places of decimal precision to display
* @return string
*/
public function getSlopeSE($dp = 0)
{
if ($dp != 0) {
return round($this->slopeSE, $dp);
}
return $this->slopeSE;
}
/**
* Return the Value of X where it intersects Y = 0
*
* @param int $dp Number of places of decimal precision to display
* @return string
*/
public function getIntersect($dp = 0)
{
if ($dp != 0) {
return round($this->intersect, $dp);
}
return $this->intersect;
}
/**
* Return the standard error of the Intersect
*
* @param int $dp Number of places of decimal precision to display
* @return string
*/
public function getIntersectSE($dp = 0)
{
if ($dp != 0) {
return round($this->intersectSE, $dp);
}
return $this->intersectSE;
}
/**
* Return the goodness of fit for this regression
*
* @param int $dp Number of places of decimal precision to return
* @return float
*/
public function getGoodnessOfFit($dp = 0)
{
if ($dp != 0) {
return round($this->goodnessOfFit, $dp);
}
return $this->goodnessOfFit;
}
public function getGoodnessOfFitPercent($dp = 0)
{
if ($dp != 0) {
return round($this->goodnessOfFit * 100, $dp);
}
return $this->goodnessOfFit * 100;
}
/**
* Return the standard deviation of the residuals for this regression
*
* @param int $dp Number of places of decimal precision to return
* @return float
*/
public function getStdevOfResiduals($dp = 0)
{
if ($dp != 0) {
return round($this->stdevOfResiduals, $dp);
}
return $this->stdevOfResiduals;
}
public function getSSRegression($dp = 0)
{
if ($dp != 0) {
return round($this->SSRegression, $dp);
}
return $this->SSRegression;
}
public function getSSResiduals($dp = 0)
{
if ($dp != 0) {
return round($this->SSResiduals, $dp);
}
return $this->SSResiduals;
}
public function getDFResiduals($dp = 0)
{
if ($dp != 0) {
return round($this->DFResiduals, $dp);
}
return $this->DFResiduals;
}
public function getF($dp = 0)
{
if ($dp != 0) {
return round($this->f, $dp);
}
return $this->f;
}
public function getCovariance($dp = 0)
{
if ($dp != 0) {
return round($this->covariance, $dp);
}
return $this->covariance;
}
public function getCorrelation($dp = 0)
{
if ($dp != 0) {
return round($this->correlation, $dp);
}
return $this->correlation;
}
public function getYBestFitValues()
{
return $this->yBestFitValues;
}
protected function calculateGoodnessOfFit($sumX, $sumY, $sumX2, $sumY2, $sumXY, $meanX, $meanY, $const)
{
$SSres = $SScov = $SScor = $SStot = $SSsex = 0.0;
foreach ($this->xValues as $xKey => $xValue) {
$bestFitY = $this->yBestFitValues[$xKey] = $this->getValueOfYForX($xValue);
$SSres += ($this->yValues[$xKey] - $bestFitY) * ($this->yValues[$xKey] - $bestFitY);
if ($const) {
$SStot += ($this->yValues[$xKey] - $meanY) * ($this->yValues[$xKey] - $meanY);
} else {
$SStot += $this->yValues[$xKey] * $this->yValues[$xKey];
}
$SScov += ($this->xValues[$xKey] - $meanX) * ($this->yValues[$xKey] - $meanY);
if ($const) {
$SSsex += ($this->xValues[$xKey] - $meanX) * ($this->xValues[$xKey] - $meanX);
} else {
$SSsex += $this->xValues[$xKey] * $this->xValues[$xKey];
}
}
$this->SSResiduals = $SSres;
$this->DFResiduals = $this->valueCount - 1 - $const;
if ($this->DFResiduals == 0.0) {
$this->stdevOfResiduals = 0.0;
} else {
$this->stdevOfResiduals = sqrt($SSres / $this->DFResiduals);
}
if (($SStot == 0.0) || ($SSres == $SStot)) {
$this->goodnessOfFit = 1;
} else {
$this->goodnessOfFit = 1 - ($SSres / $SStot);
}
$this->SSRegression = $this->goodnessOfFit * $SStot;
$this->covariance = $SScov / $this->valueCount;
$this->correlation = ($this->valueCount * $sumXY - $sumX * $sumY) / sqrt(($this->valueCount * $sumX2 - pow($sumX, 2)) * ($this->valueCount * $sumY2 - pow($sumY, 2)));
$this->slopeSE = $this->stdevOfResiduals / sqrt($SSsex);
$this->intersectSE = $this->stdevOfResiduals * sqrt(1 / ($this->valueCount - ($sumX * $sumX) / $sumX2));
if ($this->SSResiduals != 0.0) {
if ($this->DFResiduals == 0.0) {
$this->f = 0.0;
} else {
$this->f = $this->SSRegression / ($this->SSResiduals / $this->DFResiduals);
}
} else {
if ($this->DFResiduals == 0.0) {
$this->f = 0.0;
} else {
$this->f = $this->SSRegression / $this->DFResiduals;
}
}
}
protected function leastSquareFit($yValues, $xValues, $const)
{
// calculate sums
$x_sum = array_sum($xValues);
$y_sum = array_sum($yValues);
$meanX = $x_sum / $this->valueCount;
$meanY = $y_sum / $this->valueCount;
$mBase = $mDivisor = $xx_sum = $xy_sum = $yy_sum = 0.0;
for ($i = 0; $i < $this->valueCount; ++$i) {
$xy_sum += $xValues[$i] * $yValues[$i];
$xx_sum += $xValues[$i] * $xValues[$i];
$yy_sum += $yValues[$i] * $yValues[$i];
if ($const) {
$mBase += ($xValues[$i] - $meanX) * ($yValues[$i] - $meanY);
$mDivisor += ($xValues[$i] - $meanX) * ($xValues[$i] - $meanX);
} else {
$mBase += $xValues[$i] * $yValues[$i];
$mDivisor += $xValues[$i] * $xValues[$i];
}
}
// calculate slope
// $this->slope = (($this->valueCount * $xy_sum) - ($x_sum * $y_sum)) / (($this->valueCount * $xx_sum) - ($x_sum * $x_sum));
$this->slope = $mBase / $mDivisor;
// calculate intersect
// $this->intersect = ($y_sum - ($this->slope * $x_sum)) / $this->valueCount;
if ($const) {
$this->intersect = $meanY - ($this->slope * $meanX);
} else {
$this->intersect = 0;
}
$this->calculateGoodnessOfFit($x_sum, $y_sum, $xx_sum, $yy_sum, $xy_sum, $meanX, $meanY, $const);
}
/**
* Define the regression
*
* @param float[] $yValues The set of Y-values for this regression
* @param float[] $xValues The set of X-values for this regression
* @param boolean $const
*/
public function __construct($yValues, $xValues = array(), $const = true)
{
// Calculate number of points
$nY = count($yValues);
$nX = count($xValues);
// Define X Values if necessary
if ($nX == 0) {
$xValues = range(1, $nY);
$nX = $nY;
} elseif ($nY != $nX) {
// Ensure both arrays of points are the same size
$this->error = true;
return false;
}
$this->valueCount = $nY;
$this->xValues = $xValues;
$this->yValues = $yValues;
}
}