compute_clustering_accuracy.m
0fd117d9ca72master

compute_clustering_accuracy.m
View Options


			%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
			% Author:
			% Xavier Bresson, xavier.bresson@epfl.ch
			%
			% Course:
			% A Network Tour of Data Science
			% a.k.a. Graph Algorithms in Data Science
			% EPFL, Fall 2015
			%
			% Version:
			% September 23, 2015
			%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


			%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
			% Clustering accuracy can be defined with the purity measure, defined here:
			% Yang-Hao-Dikmen-Chen-Oja, Clustering by nonnegative matrix factorization
			% using graph random walk, 2012.
			%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

			%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
			% Usages:
			% accuracy = compute_clustering_accuracy(C_computed,C_grndtruth,R)
			%
			% Notations:
			% n = nb_data
			%
			% Input variables:
			% C_computed = Computed clusters. Size = n x 1. Values in [1,2,...,R].
			% C_grndtruth = Ground truth clusters. Size = n x 1. Values in [1,2,...,R].
			% R = Number of clusters.
			%
			% Output variables:
			% accuracy = Clustering accuracy of computed clusters.
			%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

			function accuracy = compute_clustering_accuracy(C_computed,C_grndtruth,R)

			n = length(C_grndtruth);
			nb_of_dominant_points_in_class = zeros(R,1);
			for k=1:R
			v = C_grndtruth(C_computed==k);
			[~,nb_of_dominant_points_in_class(k)] = mode(v);
			end
			accuracy = sum(nb_of_dominant_points_in_class)/ n;

			end