/*
 * Copyright (C) 2015 Edward Raff
 *
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * This program 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 General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see .
 */
package jsat.distributions;

import java.io.Serializable;
import java.util.Random;
import jsat.linear.DenseVector;
import jsat.linear.Vec;
import jsat.math.Function;
import jsat.math.Function1D;
import jsat.math.rootfinding.Zeroin;

/**
 * Base distribution class for distributions that have only one input. 
 * 
 * @author Edward Raff
 */
public abstract class Distribution implements Cloneable, Serializable
{

    /**
     * Computes the value of the Cumulative Density Function (CDF) at the given
     * point. The CDF returns a value in the range [0, 1], indicating what
     * portion of values occur at or below that point.
     *
     * @param x the value to get the CDF of
     * @return the CDF(x)
     */
    abstract public double cdf(double x);

    /**
     * Computes the inverse Cumulative Density Function (CDF-1) at
     * the given point. It takes in a value in the range of [0, 1] and returns
     * the value x, such that CDF(x) = p
     *
     * @param p the probability value
     * @return the value such that the CDF would return p
     */
    public double invCdf(double p)
    {
        if (p < 0 || p > 1)
            throw new ArithmeticException("Value of p must be in the range [0,1], not " + p);
        double a = Double.isInfinite(min()) ? Double.MIN_VALUE : min();
        double b = Double.isInfinite(max()) ? Double.MAX_VALUE : max();

        //default case, lets just do a root finding on the CDF for the specific value of p
        return Zeroin.root(a, b, (x) -> cdf(x) - p);
    }

    /**
     * Computes the mean value of the distribution
     *
     * @return the mean value of the distribution
     */
    abstract public double mean();

    /**
     * Computes the median value of the distribution
     *
     * @return the median value of the distribution
     */
    public double median()
    {
        return invCdf(0.5);
    }
    
    /**
     * Computes the mode of the distribution. Not all distributions have a mode for all parameter values. 
     * {@link Double#NaN NaN} may be returned if the mode is not defined for the current values of the 
     * distribution. 
     * 
     * @return the mode of the distribution
     */
    abstract public double mode();
    
    /**
     * Computes the variance of the distribution. Not all distributions have a 
     * finite variance for all parameter values. {@link Double#NaN NaN} may be 
     * returned if the variance is not defined for the current values of the distribution. 
     * {@link Double#POSITIVE_INFINITY Infinity} is a possible value to be returned
     * by some distributions. 
     * 
     * @return the variance of the distribution. 
     */
    abstract public double variance();
    
    /**
     * Computes the skewness of the distribution. Not all distributions have a 
     * finite skewness for all parameter values. {@link Double#NaN NaN} may be 
     * returned if the skewness is not defined for the current values of the distribution.
     * 
     * @return the skewness of the distribution. 
     */
    abstract public double skewness();
    
    /**
     * Computes the standard deviation of the distribution. Not all distributions have a 
     * finite standard deviation for all parameter values. {@link Double#NaN NaN} may be 
     * returned if the variance is not defined for the current values of the distribution. 
     * {@link Double#POSITIVE_INFINITY Infinity} is a possible value to be returned
     * by some distributions. 
     * 
     * @return the standard deviation of the distribution 
     */
    public double standardDeviation()
    {
        return Math.sqrt(variance());
    }

    /**
     * The minimum value for which the {@link #pdf(double) } is meant to return
     * a value. Note that {@link Double#NEGATIVE_INFINITY} is a valid return
     * value.
     *
     * @return the minimum value for which the {@link #pdf(double) } is meant to
     * return a value.
     */
    abstract public double min();

    /**
     * The maximum value for which the {@link #pdf(double) } is meant to return
     * a value. Note that {@link Double#POSITIVE_INFINITY} is a valid return
     * value.
     *
     * @return the maximum value for which the {@link #pdf(double) } is meant to
     * return a value.
     */
    abstract public double max();
    
    /**
     * This method returns a double array containing the values of random samples from this distribution. 
     * 
     * @param numSamples the number of random samples to take
     * @param rand the source of randomness
     * @return an array of the random sample values
     */
    public double[] sample(int numSamples, Random rand)
    {
        double[] samples = new double[numSamples];
        for(int i = 0; i < samples.length; i++)
            samples[i] = invCdf(rand.nextDouble());
        
        return samples;
    }
    
    /**
     * This method returns a double array containing the values of random samples from this distribution. 
     * 
     * @param numSamples the number of random samples to take
     * @param rand the source of randomness
     * @return a vector of the random sample values
     */
    public DenseVector sampleVec(int numSamples, Random rand)
    {
        return DenseVector.toDenseVec(sample(numSamples, rand));
    }
    
    @Override
    abstract public Distribution clone();
}

 

Ads help maintain this website.