Module fr.inria.MochyCore

Package fr.inria.mochy.core.sampler

Class Sampler

java.lang.Object

fr.inria.mochy.core.sampler.Sampler

public class Sampler
extends Object

Used to generate random values for different distributions

Constructor Summary

Constructors

Constructor	Description
Sampler()

Method Summary

Modifier and Type	Method	Description
float	discreteUniform(Float deb, Float end)	Sample discrete values within a given interval
double	gaussian(double x, float mean, double deviation)	return a gaussian value dependending of x, the mean, and the deviation
float	gaussSampler(float average, double deviation)
int	invertTransform(int lambda)	---(1-ln(x))²*lambda
float	invertTransformWeibull(double k, float mean)	invert repartition function of Weibull = lambda*(-log(1-x))^(1/k) repartition function of Weibull = 1-exp(-(x/lambda)^k)
double	weibull(double x, int coef, double mean)	return the weibull result of the x input with the specific parameters the coef must be > 1 to have a correct shape

Methods inherited from class java.lang.Object

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

Constructor Detail

Sampler

public Sampler()

Method Detail

discreteUniform

public float discreteUniform(Float deb,
                             Float end)

Sample discrete values within a given interval

Parameters:

deb - beginning of interval

end - end of interval

Returns:

invertTransform

public int invertTransform(int lambda)

---(1-ln(x))²*lambda

Parameters:

lambda - the parameter to increase the result

Returns:

an invertTransform value

weibull

public double weibull(double x,
                      int coef,
                      double mean)

return the weibull result of the x input with the specific parameters the coef must be > 1 to have a correct shape

invertTransformWeibull

public float invertTransformWeibull(double k,
                                    float mean)

invert repartition function of Weibull = lambda*(-log(1-x))^(1/k) repartition function of Weibull = 1-exp(-(x/lambda)^k)

Parameters:

k - the shape of the function and to increase of the result

lambda - the coefficient to increase the result

Returns:

the value

gaussian

public double gaussian(double x,
                       float mean,
                       double deviation)

return a gaussian value dependending of x, the mean, and the deviation

gaussSampler

public float gaussSampler(float average,
                          double deviation)

Parameters:

deviation - : the scale

average - : the lowerbound value

Returns:

a random value following the gaussian law