edu.washington.cs.supple.render
Class RenderedUiComparatorMatrix

java.lang.Object
  edu.washington.cs.supple.render.RenderedUiComparator
      edu.washington.cs.supple.render.RenderedUiComparatorMatrix

public class RenderedUiComparatorMatrix

extends RenderedUiComparator

Author:

anthonyw

Field Summary

Fields inherited from class edu.washington.cs.supple.render.RenderedUiComparator

function, results

Constructor Summary

RenderedUiComparatorMatrix()

Method Summary

static Matrix	calculateWeightsWithRegression(Matrix X, Matrix Y) Multivariable Linear Regression method to approximate user preference weights of UI properties
Matrix	getSimilarityCostMatrix(RenderedUi ui1, RenderedUi ui2)

Methods inherited from class edu.washington.cs.supple.render.RenderedUiComparator

areComparable, compare, getResults, getSimilarityCostFunction, setSimilarityCostFunction

Methods inherited from class java.lang.Object

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

Constructor Detail

RenderedUiComparatorMatrix

public RenderedUiComparatorMatrix()

Method Detail

getSimilarityCostMatrix

public Matrix getSimilarityCostMatrix(RenderedUi ui1,
                                      RenderedUi ui2)

Parameters:

ui1 -

ui2 -

Returns:

Matrix if compare is valid, null if ui's are not comparable

calculateWeightsWithRegression

public static Matrix calculateWeightsWithRegression(Matrix X,
                                                    Matrix Y)

Multivariable Linear Regression method to approximate user preference weights of UI properties

Parameters:

X - Set of non-weights UI difference points

Y - Difference values reported by the user

Returns:

B = (X' * X).inverse * X' * Y