lphy.base.distribution.LogNormal Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of lphy-base Show documentation
The standard library of LPhy, which contains the required generative distributions and basic functions.
The newest version!
package lphy.base.distribution;

import lphy.core.model.GenerativeDistribution1D;
import lphy.core.model.RandomVariable;
import lphy.core.model.Value;
import lphy.core.model.ValueUtils;
import lphy.core.model.annotation.GeneratorCategory;
import lphy.core.model.annotation.GeneratorInfo;
import lphy.core.model.annotation.ParameterInfo;
import org.apache.commons.math3.distribution.LogNormalDistribution;
import org.apache.commons.math3.random.RandomGenerator;

import java.util.Map;
import java.util.TreeMap;

/**
 * log-normal prior.
 * @see LogNormalDistribution
 * @author Alexei Drummond
 * @author Walter Xie
 */
public class LogNormal extends ParametricDistribution implements GenerativeDistribution1D {

    public static final String meanLogParamName = "meanlog";
    public static final String sdLogParamName = "sdlog";
    private Value M;
    private Value S;
    private Value offset;

    LogNormalDistribution logNormalDistribution;

    public LogNormal(@ParameterInfo(name = meanLogParamName, narrativeName = "mean in log space", description = "the mean of the distribution on the log scale.") Value M,
                     @ParameterInfo(name = sdLogParamName, narrativeName = "standard deviation in log space", description = "the standard deviation of the distribution on the log scale.") Value S,
                     @ParameterInfo(name = DistributionConstants.offsetParamName, optional = true, narrativeName = "offset", description = "optional parameter to shift entire distribution by an offset. default is 0.") Value offset) {
        super();
        this.M = M;
        this.S = S;
        this.offset = offset;

        constructDistribution(random);
    }

    @Override
    protected void constructDistribution(RandomGenerator random) {
        // use code available since apache math 3.1
        logNormalDistribution = new LogNormalDistribution(random, ValueUtils.doubleValue(M), ValueUtils.doubleValue(S),
                LogNormalDistribution.DEFAULT_INVERSE_ABSOLUTE_ACCURACY);
    }

    @GeneratorInfo(name = "LogNormal", verbClause = "has", narrativeName = "log-normal prior",
            category = GeneratorCategory.PRIOR, examples = {"hkyCoalescent.lphy","errorModel1.lphy"},
            description = "The log-normal probability distribution.")
    public RandomVariable sample() {
        // .sample() is before offset
        double result = logNormalDistribution.sample() + C();
        // constructDistribution() only required in constructor and setParam
        return new RandomVariable<>(null, result, this);
    }

    // default offset=0
    private double C() {
        double C = 0;
        if (offset != null) {
            C = ValueUtils.doubleValue(offset);
        }
        return C;
    }

    public double logDensity(Double x) {
        // x is after offset, so - to get the original point
        return logNormalDistribution.logDensity(x-C());
    }

    public Map getParams() {
        return new TreeMap<>() {{
            put(meanLogParamName, M);
            put(sdLogParamName, S);
        }};
    }

    public Value getMeanLog() {
        return M;
    }

    public Value getSDLog() {
        return S;
    }

    public Double[] getDomainBounds() {
        return new Double[] {C(), Double.POSITIVE_INFINITY};
    }
}