Facemorph.tensor
Class AutoRegressor

java.lang.Object
  Facemorph.tensor.AutoRegressor

public class AutoRegressor
extends java.lang.Object

Class to implement auto regression, not completed

Field Summary

static int	AR_BEST Best AR
static int	AR_MEAN Mean AR
static int	AR_ZERO Zero mean AR

Constructor Summary

AutoRegressor()

AutoRegressor(java.util.LinkedList<java.util.LinkedList<double[]>> data, int windowSize, int arConstant)
Method to calculate the regression coefficients

Method Summary

static void	add(double[] dest, double[] src, double weight) Add a weighted amount of src to dest
static double	dotProduct(double[] a, double[] b) Dot product between two vectors
double[]	predict(java.util.LinkedList<double[]> previous) Predict method
void	read(java.io.InputStream dis) Read from an InputStream
boolean	read(java.io.StreamTokenizer st) Read auto regression data
java.util.LinkedList<double[]>	test(java.util.LinkedList<double[]> original) Test method -finds the difference between the prediction and actual for each frame
java.util.LinkedList<double[]>	transform(java.util.LinkedList<double[]> original, AutoRegressor targetAr) Transform method
void	write(java.io.PrintStream ps) Write to a PrintStream

Methods inherited from class java.lang.Object

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

Field Detail

AR_ZERO

public static final int AR_ZERO

Zero mean AR

See Also:

Constant Field Values

AR_MEAN

public static final int AR_MEAN

Mean AR

See Also:

Constant Field Values

AR_BEST

public static final int AR_BEST

Best AR

See Also:

Constant Field Values

Constructor Detail

AutoRegressor

public AutoRegressor()

AutoRegressor

public AutoRegressor(java.util.LinkedList<java.util.LinkedList<double[]>> data,
                     int windowSize,
                     int arConstant)

Method to calculate the regression coefficients

Parameters:

data - The list of data vectors in order

windowSize - The size of the window to use in the regression

arConstant - indicates whether to use zero, the mean or the best constant vector

Method Detail

write

public void write(java.io.PrintStream ps)

Write to a PrintStream

Parameters:

ps - the PrintStream to write to

read

public void read(java.io.InputStream dis)

Read from an InputStream

Parameters:

dis - the InputStream to read from

read

public boolean read(java.io.StreamTokenizer st)

Read auto regression data

Parameters:

st - the StreamTokenizer to read from

Returns:

returns true if read OK

dotProduct

public static double dotProduct(double[] a,
                                double[] b)

Dot product between two vectors

Parameters:

a - the first vector

b - the second vector

Returns:

return the dot product

add

public static void add(double[] dest,
                       double[] src,
                       double weight)

Add a weighted amount of src to dest

Parameters:

dest - the dest vector

src - the src vector

weight - the amount of src to add to dest

predict

public double[] predict(java.util.LinkedList<double[]> previous)

Predict method

Parameters:

previous - set of vectors

Returns:

returns a weighted sum of the previous

transform

public java.util.LinkedList<double[]> transform(java.util.LinkedList<double[]> original,
                                                AutoRegressor targetAr)

Transform method

Parameters:

original - the original vector

targetAr - the auto regression

Returns:

return the transformed data

test

public java.util.LinkedList<double[]> test(java.util.LinkedList<double[]> original)

Test method -finds the difference between the prediction and actual for each frame

Parameters:

original - the original vector

Returns:

return the residual error data