edu.stanford.rsl.conrad.utils

Class StatisticsUtil

java.lang.Object

edu.stanford.rsl.conrad.utils.StatisticsUtil

public abstract class StatisticsUtil
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
StatisticsUtil()

Method Summary

Methods

Modifier and Type	Method and Description
static double	logfactorial(int k) Computes ln(k!) using the Stirling approximation.
static void	main(java.lang.String[] args)
static int	poissonRandomNumber(double lambda) Generates a Poisson distributed random number.
static int	poissonRandomNumberBig(double lambda) Generates a Poisson distributed random number.
static int	poissonRandomNumberSmall(double lambda) Generates a Poisson distributed random number.

Methods inherited from class java.lang.Object

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

Constructor Detail

StatisticsUtil

public StatisticsUtil()

Method Detail

poissonRandomNumber

public static int poissonRandomNumber(double lambda)

Generates a Poisson distributed random number. The Poisson distribution is definded by the number lambda. The distribution has mean lambda and standard deviation of Math.sqrt(lambda).
If lambda is smaller than 200 the value is determined by cumulative statistics, otherwise an approximation is used. Time measurements showed that the approximation used here outperforms the implementation in Matlab 2008.

Parameters:

lambda -

Returns:

a random number drawn from the Poisson distribution

See Also:

poissonRandomNumberSmall(double), poissonRandomNumberBig(double)

poissonRandomNumberSmall

public static int poissonRandomNumberSmall(double lambda)

Generates a Poisson distributed random number. The Poisson distribution is definded by the number lambda. The distribution has mean lambda and standard deviation of Math.sqrt(lambda).
Computation is based on cumulative statistics and gets slow with greater values of lambda.

Parameters:

lambda -

Returns:

a random number from the Poisson distribution

logfactorial

public static double logfactorial(int k)

Computes ln(k!) using the Stirling approximation.
(cf. J.H. Ahrens, U. Dieter. Computer Generation of Poission Deviates from Modified Normal Distributions. ACM Transactions on Mathematical Software (TOMS). 8(2):163-79. 1982.)

Parameters:

k - the k

Returns:

ln(k!)

poissonRandomNumberBig

public static int poissonRandomNumberBig(double lambda)

Generates a Poisson distributed random number. The Poisson distribution is definded by the number lambda. The distribution has mean lambda and standard deviation of Math.sqrt(lambda).
Uses the rejection algorithm after Atkinson (cf. A.C. Atkinson. The Computer Generation of Poisson Random Variables. Journal of the Royal Statistical Society. Series C (Applied Statistics) Vol. 28, No. 1 (1979), pp. 29-35)

Parameters:

lambda -

Returns:

a random number from the Poisson distribution

main

public static void main(java.lang.String[] args)