Package io.bhagat.ai.unsupervised

Class PrincipalComponentAnalysis

java.lang.Object

io.bhagat.ai.unsupervised.PrincipalComponentAnalysis

All Implemented Interfaces:

java.io.Serializable

public class PrincipalComponentAnalysis
extends java.lang.Object
implements java.io.Serializable

A class containing methods for Principal Component Analysis

Author:

Bhagat

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
private Matrix	C
int	curNumOfDimensions	the current number of dimensions
Matrix	data
double	defaultThreshold	the default threshold to decide whether a component has enough variance
private double[]	eigenvalues
private Vector[]	eigenvectors
private static long	serialVersionUID
private Matrix	X

Constructor Summary

Constructors

Constructor	Description
PrincipalComponentAnalysis(Matrix X)	Creates an object that will perform principal component analysis

Method Summary

Modifier and Type	Method	Description
static Matrix	centerData(Matrix X)	Centers the data from the matrix X at the column mean
static Matrix	covarianceMatrix(Matrix X)	creates the covariance matrix for the data
static Matrix	createMatrix(byte[] data)	creates a matrix from a byte array
Matrix	dimensionReduction()	Reduces the dimensions of the data matrix so only the important dimensions remain
Matrix	dimensionReduction(double threshold)	Reduces the dimensions of the data matrix so only the important dimensions remain
Matrix	dimensionReduction(int D)	Reduces the dimensions of the data matrix into D dimensions
Matrix	dimensionReduction(Matrix X)	Reduces the dimensions of the data matrix so only the important dimensions remain
Matrix	dimensionReduction(Matrix X, double threshhold)	Reduces the dimensions of the data matrix so only the important dimensions remain
Matrix	dimensionReduction(Matrix X, int D)	Reduces the dimensions of the data matrix into D dimensions
Matrix	dimensionReduction(Vector x)	Reduces the dimensions of the data point into D dimensions
Matrix	dimensionReduction(Vector x, double threshold)	Reduces the dimensions of the data point into D dimensions
Matrix	dimensionReduction(Vector x, int D)	Reduces the dimensions of the data point into D dimensions
Matrix	getCenteredData()
void	setDataMatrix(Matrix X)	sets the data of the object

Methods inherited from class java.lang.Object

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

Field Detail

serialVersionUID

private static final long serialVersionUID

See Also:

Constant Field Values

defaultThreshold

public double defaultThreshold

the default threshold to decide whether a component has enough variance

curNumOfDimensions

public int curNumOfDimensions

the current number of dimensions

data

public Matrix data

X

private Matrix X

C

private Matrix C

eigenvalues

private double[] eigenvalues

eigenvectors

private Vector[] eigenvectors

Constructor Detail

PrincipalComponentAnalysis

public PrincipalComponentAnalysis(Matrix X)

Creates an object that will perform principal component analysis

Parameters:

X - the data matrix

Method Detail

setDataMatrix

public void setDataMatrix(Matrix X)

sets the data of the object

Parameters:

X - the data matrix

getCenteredData

public Matrix getCenteredData()

Returns:

gets the standardized data matrix X

createMatrix

public static Matrix createMatrix(byte[] data)

creates a matrix from a byte array

Parameters:

data - the byte data

Returns:

the matrix

dimensionReduction

public Matrix dimensionReduction(Matrix X,
                                 int D)

Reduces the dimensions of the data matrix into D dimensions

Parameters:

X - the data to reduce the dimensions of

D - the number of ending dimensions

Returns:

the transformed data

dimensionReduction

public Matrix dimensionReduction(Matrix X,
                                 double threshhold)

Reduces the dimensions of the data matrix so only the important dimensions remain

Parameters:

X - the data matrix

threshhold - the threshold for the eigenvalues defining what is important

Returns:

the reduced data matrix

dimensionReduction

public Matrix dimensionReduction(Matrix X)

Reduces the dimensions of the data matrix so only the important dimensions remain

Parameters:

X - the data matrix

Returns:

the reduced data matrix

dimensionReduction

public Matrix dimensionReduction(int D)

Reduces the dimensions of the data matrix into D dimensions

Parameters:

D - the number of ending dimensions

Returns:

the transformed data

dimensionReduction

public Matrix dimensionReduction(double threshold)

Reduces the dimensions of the data matrix so only the important dimensions remain

Parameters:

threshold - the threshold for the eigenvalues defining what is important

Returns:

the reduced data matrix

dimensionReduction

public Matrix dimensionReduction()

Reduces the dimensions of the data matrix so only the important dimensions remain

Returns:

the reduced data matrix

dimensionReduction

public Matrix dimensionReduction(Vector x,
                                 int D)

Reduces the dimensions of the data point into D dimensions

Parameters:

x - a vector data point

D - the number of ending dimensions

Returns:

the transformed data

dimensionReduction

public Matrix dimensionReduction(Vector x,
                                 double threshold)

Reduces the dimensions of the data point into D dimensions

Parameters:

x - a vector data point

threshold - the threshold for the eigenvalues defining what is important

Returns:

the transformed data

dimensionReduction

public Matrix dimensionReduction(Vector x)

Reduces the dimensions of the data point into D dimensions

Parameters:

x - a vector data point

Returns:

the transformed data

centerData

public static Matrix centerData(Matrix X)

Centers the data from the matrix X at the column mean

Parameters:

X - the data

Returns:

the centered data

covarianceMatrix

public static Matrix covarianceMatrix(Matrix X)

creates the covariance matrix for the data

Parameters:

X - the data matrix

Returns:

the covariance matrix