metdetection.utils

Class SmoothingUtils

java.lang.Object

metdetection.utils.SmoothingUtils

public class SmoothingUtils
extends java.lang.Object

A static class of some static utility functions that perform Good-Turing smoothing as in Church et al (1991).

Author:

Andrew McLeod - 25 March, 2016

Constructor Summary

Constructors

Modifier	Constructor and Description
private	SmoothingUtils() A private constructor since this class should never be instantiated.

Method Summary

Methods

Modifier and Type	Method and Description
static void	addFrequency(java.util.Map<java.lang.Integer,java.lang.Integer> frequencyCounts, int frequency) Add one to the count of the given frequency in the given Map.
private static double	getFrequencyCountEstimate(double[] regression, int frequency) Get the frequency count estimate of the given frequency based on the values from the regression from a log-log plot.
static java.util.Map<java.lang.Integer,java.lang.Integer>	getFrequencyMap(java.util.Collection<java.lang.Integer> counts) Get the frequency counts of the counts in the given Collection.
static java.util.Map<java.lang.Integer,java.lang.Double>	getGoodTuringSmoothing(java.util.Map<java.lang.Integer,java.lang.Integer> frequencyCounts, int count) Return the Good-Turing smoothed probability values for our counts.
static int	getTotalCount(java.util.Collection<java.lang.Integer> counts) Get the sum of the counts in the given collection.
private static double[]	linearRegression(double[] x, double[] y) Perform linear regression given x and y coordinates.
static void	removeFrequency(java.util.Map<java.lang.Integer,java.lang.Integer> frequencyCounts, int frequency) Remove one count from the given frequency in the given Map.

Methods inherited from class java.lang.Object

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

Constructor Detail

SmoothingUtils

private SmoothingUtils()

A private constructor since this class should never be instantiated.

Method Detail

getGoodTuringSmoothing

public static java.util.Map<java.lang.Integer,java.lang.Double> getGoodTuringSmoothing(java.util.Map<java.lang.Integer,java.lang.Integer> frequencyCounts,
                                                                       int count)

Return the Good-Turing smoothed probability values for our counts.

Parameters:

frequencyCounts - A Map of the frequency of an object to the number of objects with that frequency.

count - The total count of all objects.

Returns:

A Map of the Good-Turing smoothed probability values for each frequency.

getFrequencyCountEstimate

private static double getFrequencyCountEstimate(double[] regression,
                               int frequency)

Get the frequency count estimate of the given frequency based on the values from the regression from a log-log plot.

Parameters:

regression - The slope (index 0) and intercept (index 1) of the regression line (from a log-log plot).

frequency - The frequency whose count we want.

Returns:

The count of the given frequency from the regression line.

linearRegression

private static double[] linearRegression(double[] x,
                        double[] y)

Perform linear regression given x and y coordinates.
This algorithm was adapted from http://introcs.cs.princeton.edu/java/97data/LinearRegression.java.html

Parameters:

x - The x coordinates of our points.

y - The y coordinates of our points.

Returns:

A length 2 array containing the slope at index 0 and the intercept at index 1.

removeFrequency

public static void removeFrequency(java.util.Map<java.lang.Integer,java.lang.Integer> frequencyCounts,
                   int frequency)

Remove one count from the given frequency in the given Map.

Parameters:

frequencyCounts - A Map of the frequency of an object to the number of objects with that frequency. We will decrement 1 from this Map.

frequency - The frequency whose count we want to decrement.

addFrequency

public static void addFrequency(java.util.Map<java.lang.Integer,java.lang.Integer> frequencyCounts,
                int frequency)

Add one to the count of the given frequency in the given Map.

Parameters:

frequencyCounts - A Map of the frequency of an object to the number of objects with that frequency. We will increment one in this Map.

frequency - The frequency whose count we want to increment.

getFrequencyMap

public static java.util.Map<java.lang.Integer,java.lang.Integer> getFrequencyMap(java.util.Collection<java.lang.Integer> counts)

Get the frequency counts of the counts in the given Collection.

Parameters:

counts - The Collection whose frequency counts we want.

Returns:

A Map of a count to the frequency of that count in the given Collection.

getTotalCount

public static int getTotalCount(java.util.Collection<java.lang.Integer> counts)

Get the sum of the counts in the given collection.

Parameters:

counts - A Collection of counts which we want to sum.

Returns:

A sum of the counts in the given Collection.