edu.sc.seis.fissuresUtil.bag
Class Statistics

java.lang.Object
  edu.sc.seis.fissuresUtil.bag.Statistics

public class Statistics
extends java.lang.Object

Statistics.java Created: Wed Apr 4 22:27:52 2001

Field Summary

protected double[]	autocorrelation
protected int	beginIndex
protected double[]	dSeries
protected int	endIndex
protected float[]	fSeries
protected int[]	iSeries
protected double[]	minMaxMean index 0 has min, index 1 has max, index 2 has mean.
protected boolean	minMaxMeanCalculated Has min/max/mean already been calculated and cached.
protected double[]	partialautocorr
protected short[]	sSeries
protected double	variance
protected boolean	varianceCalculated Has the variable been calculated and cached.

Constructor Summary

Statistics(double[] dSeries)
Creates a new Statistics instance.

Statistics(float[] fSeries)
Creates a new Statistics instance.

Statistics(int[] iSeries)
Creates a new Statistics instance.

Statistics(java.util.List<QuantityImpl> vals)

Statistics(java.util.List<QuantityImpl> vals, UnitImpl unit)

Statistics(LocalSeismogramImpl seismo)
Creates a new Statistics instance.

Statistics(short[] sSeries)
Creates a new Statistics instance.

Method Summary

double[]	acf(int maxlag) Calculates the autocovariance function out to the given lag.
double[]	acf95conf(int maxlag) Describe acf95conf method here.
double[]	acfTRatio(int maxlag) Calculates the TRatio for the acf out to the given max lag.
double	binaryIndexSum(int start, int finish) Calulates the sum(i*yi) from beginIndex to endIndex-1.
double	binarySum(int start, int finish) Calulates the sum of the series from beginIndex to endIndex-1.
double	binarySumDevLag(int start, int finish, double mean, int lag) Describe binarySumDevLag method here.
double	binarySumDevSqr(int start, int finish, double mean) Calulates the sum of the square of the difference from the mean.
double	correlation(double[] other)
double	covariance(double[] other)
double	covariance(Statistics otherStat)
int	getLength() The length of the series.
int[]	histogram(double start, double width, int number) Creates a histogram of the values.
double[]	linearLeastSquares() Calculates the linear Least Squares slope and intercept for this series.
double	max() Finds the max value.
double	max(int beginIndex, int endIndex) Finds the max value between index beginIndex and endIndex-1.
double	maxDeviation() Finds the max deviation from the mean.
double	maxDeviation(int beginIndex, int endIndex) Finds the max deviation from the mean between index beginIndex and endIndex-1.
double	mean() Calulates mean of the data.
double	mean(int beginIndex, int endIndex) Finds the mean value between index beginIndex and endIndex-1.
double	min() Finds the min value.
double	min(int beginIndex, int endIndex) Finds the min value between index beginIndex and endIndex-1.
double[]	minMaxMean() Calculates the min, max and mean.
double[]	minMaxMean(int beginIndex, int endIndex) Calculates the min, max and mean from beginIndex to endIndex-1.
double[]	pacf(int maxlag) Computes the partial autocorrelation function, after Wei, William S.
double	pacf95conf(int maxlag) Calculates the 95 percent confidence for the pacf.
double[]	pacfTRatio(int maxlag) Calculates the TRatio for the pacf.
double	rms()
double	stddev() Calulates the standard deviation.
java.lang.String	toString()
double	var() Calualates the unbiased variance, ie 1/(n-1) SUM{(x-mean)^2}.
double	var(double mean) calculates the unbiased variance w.r.t the given mean.

Methods inherited from class java.lang.Object

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

Field Detail

iSeries

protected int[] iSeries

sSeries

protected short[] sSeries

fSeries

protected float[] fSeries

dSeries

protected double[] dSeries

minMaxMeanCalculated

protected boolean minMaxMeanCalculated

Has min/max/mean already been calculated and cached.

minMaxMean

protected double[] minMaxMean

index 0 has min, index 1 has max, index 2 has mean.

variance

protected double variance

varianceCalculated

protected boolean varianceCalculated

Has the variable been calculated and cached.

autocorrelation

protected double[] autocorrelation

partialautocorr

protected double[] partialautocorr

beginIndex

protected int beginIndex

endIndex

protected int endIndex

Constructor Detail

Statistics

public Statistics(int[] iSeries)

Creates a new Statistics instance.

Parameters:

iSeries - an int[] value

Statistics

public Statistics(short[] sSeries)

Creates a new Statistics instance.

Parameters:

sSeries - a short[] value

Statistics

public Statistics(float[] fSeries)

Creates a new Statistics instance.

Parameters:

fSeries - a float[] value

Statistics

public Statistics(double[] dSeries)

Creates a new Statistics instance.

Parameters:

dSeries - a double[] value

Statistics

public Statistics(LocalSeismogramImpl seismo)
           throws FissuresException

Creates a new Statistics instance.

Parameters:

seismo - a LocalSeismogramImpl value

Throws:

FissuresException

Statistics

public Statistics(java.util.List<QuantityImpl> vals)

Statistics

public Statistics(java.util.List<QuantityImpl> vals,
                  UnitImpl unit)

Method Detail

min

public double min()

Finds the min value.

Returns:

the minimum

min

public double min(int beginIndex,
                  int endIndex)

Finds the min value between index beginIndex and endIndex-1.

Parameters:

beginIndex - first index to search

endIndex - end index, last + 1

Returns:

a double value

max

public double max()

Finds the max value.

Returns:

a double value

max

public double max(int beginIndex,
                  int endIndex)

Finds the max value between index beginIndex and endIndex-1.

Parameters:

beginIndex - an int value

endIndex - an int value

Returns:

a double value

maxDeviation

public double maxDeviation()

Finds the max deviation from the mean.

Returns:

a double value

maxDeviation

public double maxDeviation(int beginIndex,
                           int endIndex)

Finds the max deviation from the mean between index beginIndex and endIndex-1.

Parameters:

beginIndex - an int value

endIndex - an int value

Returns:

a double value

mean

public double mean()

Calulates mean of the data.

Returns:

a double value

mean

public double mean(int beginIndex,
                   int endIndex)

Finds the mean value between index beginIndex and endIndex-1.

Parameters:

beginIndex - an int value

endIndex - an int value

Returns:

a double value

minMaxMean

public double[] minMaxMean()

Calculates the min, max and mean. The value at index 0 is the min, at index 1 is the max and at index 2 is the mean

Returns:

min, max, mean in a double[3]

minMaxMean

public double[] minMaxMean(int beginIndex,
                           int endIndex)

Calculates the min, max and mean from beginIndex to endIndex-1. The value at index 0 is the min, at index 1 is the max and at index 2 is the mean

Parameters:

beginIndex - first index

endIndex - last index +1

Returns:

min, max, mean in a double[3]

var

public double var()

Calualates the unbiased variance, ie 1/(n-1) SUM{(x-mean)^2}.

Returns:

variance

var

public double var(double mean)

calculates the unbiased variance w.r.t the given mean. This allows a psuedo-variance to be calculated based on a longer term mean.

rms

public double rms()

stddev

public double stddev()

Calulates the standard deviation. This is a shortcut for Math.sqrt(var()); Note this uses the unbiased variance.

Returns:

the Standard deviation

covariance

public double covariance(double[] other)

covariance

public double covariance(Statistics otherStat)

correlation

public double correlation(double[] other)

linearLeastSquares

public double[] linearLeastSquares()

Calculates the linear Least Squares slope and intercept for this series. Note that this is calculated relative to the index of the array, with zero index as the first data point.

Returns:

the intercept in index 0 and the slope in index 1

acf

public double[] acf(int maxlag)

Calculates the autocovariance function out to the given lag.

Parameters:

maxlag - the maximum lag to calculate the acf

Returns:

the acf for the series out to maxlag

acf95conf

public double[] acf95conf(int maxlag)

Describe acf95conf method here.

Parameters:

maxlag - an int value

Returns:

a double[] value

acfTRatio

public double[] acfTRatio(int maxlag)

Calculates the TRatio for the acf out to the given max lag.

Parameters:

maxlag - the maximu lag

Returns:

the T ratio

pacf

public double[] pacf(int maxlag)

Computes the partial autocorrelation function, after Wei, William S. Time Series Analysis, pp 22-23.

Parameters:

maxlag - the maximum lag

Returns:

the pacf

pacf95conf

public double pacf95conf(int maxlag)

Calculates the 95 percent confidence for the pacf. This is just 1.96*sqrt(length)

Parameters:

maxlag - the maximum lag

Returns:

the 95 percent conf for the pacf

pacfTRatio

public double[] pacfTRatio(int maxlag)

Calculates the TRatio for the pacf.

Parameters:

maxlag - the maximum lag

Returns:

T ratio for the pacf

getLength

public int getLength()

The length of the series.

histogram

public int[] histogram(double start,
                       double width,
                       int number)

Creates a histogram of the values. Each value is added to the bin Math.floor((value-start)/width) and the returned int array has length number

Parameters:

start - binning start value

width - bin width

number - number of bins

Returns:

an histogram

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

binarySum

public double binarySum(int start,
                        int finish)

Calulates the sum of the series from beginIndex to endIndex-1. This is done recursively in halves to avoid rounding errors.

Parameters:

start - starting index

finish - last index +1

Returns:

the sum

binarySumDevSqr

public double binarySumDevSqr(int start,
                              int finish,
                              double mean)

Calulates the sum of the square of the difference from the mean. Sum ((xi-mean)^2)

Parameters:

start - first index

finish - last index+1

mean - the mean

Returns:

the result

binarySumDevLag

public double binarySumDevLag(int start,
                              int finish,
                              double mean,
                              int lag)

Describe binarySumDevLag method here.

Parameters:

start - an int value

finish - an int value

mean - a double value

lag - an int value

Returns:

a double value

binaryIndexSum

public double binaryIndexSum(int start,
                             int finish)

Calulates the sum(i*yi) from beginIndex to endIndex-1. This is done recursively in halves to avoid rounding errors. This is mainly used as one of the terms in a linear least squares.

Parameters:

start - starting index

finish - last index +1

Returns:

the sum