math.stats.distribution.AbstractContinuousDistribution Maven / Gradle / Ivy
Show all versions of finwhale Show documentation
/*
* Copyright 2013 SPZ
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package math.stats.distribution;
import math.rng.PseudoRandom;
public abstract class AbstractContinuousDistribution implements
ContinuousDistribution {
protected final PseudoRandom prng;
protected AbstractContinuousDistribution(final PseudoRandom prng) {
this.prng = prng;
}
@Override
public double probability(final double x0, final double x1) {
if (x0 > x1) {
throw new IllegalArgumentException("Lower endpoint (" + x0
+ ") must be less than or equal to upper endpoint (" + x1
+ ")");
}
return cdf(x1) - cdf(x0);
}
/**
* {@inheritDoc}
*/
@Override
public double[] sample(final int sampleSize) {
final double[] samples = new double[sampleSize];
for (int i = 0; i < sampleSize; ++i) {
samples[i] = sample();
}
return samples;
}
protected final String getSimpleName(Number... params) {
String name = getClass().getSimpleName();
StringBuilder buf = new StringBuilder(name);
buf.append("(");
for (int i = 0; i < params.length; ++i) {
buf.append(params[i]);
if (i != params.length - 1) {
buf.append(", ");
}
}
buf.append(")");
return buf.toString();
}
private static final double FINDROOT_ACCURACY = 1.0e-15;
private static final int FINDROOT_MAX_ITERATIONS = 150;
/**
* This method approximates the value of {@code x} for which
* {@code P(X <= x) = p} where {@code p} is a given probability.
*
* It applies a combination of the Newton-Raphson algorithm and the
* bisection method to the value {@code start} as a starting point.
*
* Furthermore, to ensure convergence and stability, the caller must supply
* an interval {@code [xMin, xMax]} in which the probability distribution
* reaches the value {@code p}.
*
* Caution: this method does not check its arguments! It will produce wrong
* results if bad values for the parameters are supplied. To be used with
* care!
*
* @param p
* the given probability for which we want to find the
* corresponding value of {@code x} such that
* {@code P(X <= x) = p}
* @param start
* an initial guess that must lie in the interval
* {@code [xMin, xMax]} as a starting point for the search for
* {@code x}
* @param xMin
* lower bound for an interval that must contain the searched
* {@code x}
* @param xMax
* upper bound for an interval that must contain the searched
* {@code x}
* @return an approximation for the value of {@code x} for which
* {@code P(X <= x) = p}
*/
protected final double findRoot(double p, double start, double xMin,
double xMax) {
double x = start;
double xNew = start;
double dx = 1.0;
int i = 0;
while (Math.abs(dx) > FINDROOT_ACCURACY
&& i++ < FINDROOT_MAX_ITERATIONS) {
// apply Newton-Raphson step
double error = cdf(x) - p;
if (error < 0.0) {
xMin = x;
} else {
xMax = x;
}
double density = pdf(x);
if (density != 0.0) { // avoid division by zero
dx = error / density;
xNew = x - dx;
}
// If Newton-Raphson fails to converge (which, for example, may be
// the case if the initial guess is too rough) we apply a bisection
// step to determine a more narrow interval around the root
if (xNew < xMin || xNew > xMax || density == 0.0) {
xNew = (xMin + xMax) / 2.0;
dx = xNew - x;
}
x = xNew;
}
return x;
}
}