/**
* __construct Function
*
* Constructor function for this regression model. Takes two arrays
* of data that are parallel arrays of independent and dependent
* observations.
*
* @param array $datax The series of independent variables
* @param array $datay The series of dependent variables
* @return LogarithmicRegression An object representing the regression model
*/
public function __construct($datax, $datay)
{
$logx = array();
foreach ($datax as $x) {
$logx[] = log($x);
}
$this->r = \PHPStats\Stats::correlation($logx, $datay);
$this->beta = \PHPStats\Stats::covariance($logx, $datay) / \PHPStats\Stats::variance($logx);
$this->alpha = \PHPStats\Stats::average($datay) - $this->beta * \PHPStats\Stats::average($logx);
}
/**
* __construct Function
*
* Constructor function for this regression model. Takes two arrays
* of data that are parallel arrays of independent and dependent
* observations.
*
* @param array $datax The series of independent variables
* @param array $datay The series of dependent variables
* @return ExponentialRegression An object representing the regression model
*/
public function __construct(array $datax, array $datay)
{
$logy = array();
foreach ($datay as $y) {
$logy[] = log($y);
}
$this->r = \PHPStats\Stats::correlation($datax, $logy);
$logbeta = \PHPStats\Stats::covariance($datax, $logy) / \PHPStats\Stats::variance($datax);
$logalpha = \PHPStats\Stats::average($logy) - $logbeta * \PHPStats\Stats::average($datax);
$this->beta = exp($logbeta);
$this->alpha = exp($logalpha);
}
/**
Returns the cumulative distribution function, the probability of getting the test value or something below it
@param float $x The test value
@param int $L The population size
@param int $m The number of interesting elements in the population
@param int $n The number of draws from the population
@return float The probability
*/
static function getCdf($x, $L = 1, $m = 1, $n = 1)
{
$x = floor($x);
$L = floor($L);
$m = floor($m);
$n = floor($n);
if ($L >= 1 && $m >= 0 && $n >= 0) {
$sum = 0;
for ($i = 0; $i <= $x; $i++) {
$sum += \PHPStats\Stats::combinations($m, $i) * \PHPStats\Stats::combinations($L - $m, $n - $i) / \PHPStats\Stats::combinations($L, $n);
}
return $sum;
} else {
return 0.0;
}
}
/**
Returns the cumulative distribution function, the probability of getting the test value or something below it
@param float $x The test value
@param float $k Shape parameter
@param float $theta Scale parameter
@return float The probability
*/
static function getCdf($x, $k = 1, $theta = 1)
{
return \PHPStats\Stats::lowerGamma($k, $x / $theta) / \PHPStats\Stats::gamma($k);
}
/**
Returns the cumulative distribution function, the probability of getting the test value or something below it
@param float $x The test value
@param float $k Shape parameter
@return float The probability
*/
static function getCdf($x, $k = 1)
{
return \PHPStats\Stats::lowerGamma($k / 2.0, $x / 2) / \PHPStats\Stats::gamma($k / 2.0);
}
/**
Returns the cumulative distribution function, the probability of getting the test value or something below it
@param float $x The test value
@param float $mu The location parameter. Default 0.0
@param float $variance The scale parameter. Default 1.0
@return float The probability
*/
static function getCdf($x, $mu = 0.0, $variance = 1.0)
{
return (1 + \PHPStats\Stats::erf(($x - $mu) / sqrt(2 * $variance))) / 2;
}
/**
Returns the probability mass function
@param float $x The test value
@param float $lambda The rate of events
@return float The probability
*/
static function getPmf($x, $lambda = 1)
{
return exp(-$lambda) * pow($lambda, $x) / \PHPStats\Stats::factorial($x);
}
/**
* Returns the probability mass function
*
* @param float $x The test value
* @param float $p The probability of success per trial
* @param int $n The number of trials
* @return float The probability
*/
public static function getPmf($x, $p = 0.5, $n = 1)
{
return \PHPStats\Stats::combinations($n, $x) * pow($p, $x) * pow(1 - $p, $n - $x);
}
/**
* __construct Function
*
* Constructor function for this regression model. Takes two arrays
* of data that are parallel arrays of independent and dependent
* observations.
*
* @param array $datax The series of independent variables
* @param array $datay The series of dependent variables
* @return LinearRegression An object representing the regression model
*/
public function __construct(array $datax, array $datay)
{
$this->beta = \PHPStats\Stats::covariance($datax, $datay) / \PHPStats\Stats::variance($datax);
$this->alpha = \PHPStats\Stats::average($datay) - $this->beta * \PHPStats\Stats::average($datax);
$this->r = \PHPStats\Stats::correlation($datax, $datay);
}
/**
Returns the cumulative distribution function, the probability of getting the test value or something below it
@param float $x The test value
@param float $alpha The minimum parameter. Default 0.0
@param float $beta The maximum parameter. Default 1.0
@return float The probability
*/
static function getCdf($x, $alpha = 1, $beta = 1)
{
return \PHPStats\Stats::regularizedIncompleteBeta($alpha, $beta, $x);
}
/**
Returns the cumulative distribution function, the probability of getting the test value or something below it
@param float $x The test value
@param float $df The degrees of freedeom. Default 1
@return float The probability
*/
static function getCdf($x, $df = 1)
{
$return = 1 - 0.5 * \PHPStats\Stats::regularizedIncompleteBeta($df / 2, 0.5, $df / (pow($x, 2) + $df));
//Valid only for $x > 0
if ($x < 0) {
return 1 - $return;
} elseif ($x == 0) {
return 0.5;
} else {
return $return;
}
}
public function test_combinations()
{
$this->assertEquals(1, Stats::combinations(1, 1));
$this->assertEquals(2, Stats::combinations(2, 1));
$this->assertEquals(6, Stats::combinations(4, 2));
$this->assertEquals(1, Stats::combinations(5, 5));
$this->assertEquals(56, Stats::combinations(8, 5));
}