/*
* Concept profile generation tool suite
* Copyright (C) 2015 Biosemantics Group, Erasmus University Medical Center,
* Rotterdam, The Netherlands
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published
* by the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>
*/
package org.erasmusmc.math;
public class Kappa {
public int[][] ratingMatrix;
public double kappa;
public double variance;
public int numberOfObjects;
public int numberOfCategories;
public int numberOfClassifiers;
/**
* zScore only valid with larger N (about >20)
*/
public double zScore;
/**
* This implementation expects a rating Matrix, where the rows (first index)
* represent the objects to be classified and the columns (second index) the
* classes. For every row the number of columns should be equal. The sum of
* the row values should be equal to the number of classifiers k for every
* row. If you do not know what the beforementioned means: RAFM on the kappa
* statistic, thank you. Cheers Rob.
*/
public Kappa(int[][] ratingMatrix) {
numberOfCategories = ratingMatrix[0].length;
int total = 0;
int summedSquaredCellValues = 0;
numberOfClassifiers = 0;
numberOfObjects = ratingMatrix.length;
int[] categories = new int[numberOfCategories];
for (int i = 0; i < ratingMatrix.length; i++) {
for (int j = 0; j < numberOfCategories; j++) {
int cellValue = ratingMatrix[i][j];
categories[j] += cellValue;
summedSquaredCellValues += cellValue * cellValue;
}
}
for (int j = 0; j < numberOfCategories; j++) {
total += categories[j];
numberOfClassifiers += ratingMatrix[0][j];
}
double expectedAgreement = 0d;
double pj3 = 0d;
for (int j = 0; j < numberOfCategories; j++) {
Double chanceForCategoryJ = ((double) categories[j]) / ((double) total);
expectedAgreement += chanceForCategoryJ * chanceForCategoryJ;
pj3 += Math.pow(chanceForCategoryJ, 3d);
}
Double proportionOfAgreement = -1d / (double) (numberOfClassifiers - 1) + (double) summedSquaredCellValues / (double) (total * (numberOfClassifiers - 1));
if (total == numberOfClassifiers * ratingMatrix.length) {
kappa = (proportionOfAgreement - expectedAgreement) / (1d - expectedAgreement);
variance = 2d / (double) (numberOfObjects * numberOfClassifiers * (numberOfClassifiers - 1));
variance *= expectedAgreement - (2d * numberOfClassifiers - 3d) * expectedAgreement * expectedAgreement + 2 * (numberOfClassifiers - 2) * pj3;
variance /= (1d - expectedAgreement) * (1d - expectedAgreement);
zScore = kappa / Math.sqrt(variance);
}
else {
kappa = -1d;
}
}
}