// Copyright (C) 2012 Davis E. King (davis@dlib.net)
// License: Boost Software License See LICENSE.txt for the full license.
#include <dlib/clustering.h>
#include "tester.h"
namespace
{
using namespace test;
using namespace dlib;
using namespace std;
logger dlog("test.clustering");
// ----------------------------------------------------------------------------------------
void make_test_graph(
dlib::rand& rnd,
std::vector<sample_pair>& edges,
std::vector<unsigned long>& labels,
const int groups,
const int group_size,
const int noise_level,
const double missed_edges
)
{
labels.resize(groups*group_size);
for (unsigned long i = 0; i < labels.size(); ++i)
{
labels[i] = i/group_size;
}
edges.clear();
for (int i = 0; i < groups; ++i)
{
for (int j = 0; j < group_size; ++j)
{
for (int k = 0; k < group_size; ++k)
{
if (j == k)
continue;
if (rnd.get_random_double() < missed_edges)
continue;
edges.push_back(sample_pair(j+group_size*i, k+group_size*i, 1));
}
}
}
for (int k = 0; k < groups*noise_level; ++k)
{
const int i = rnd.get_random_32bit_number()%labels.size();
const int j = rnd.get_random_32bit_number()%labels.size();
edges.push_back(sample_pair(i,j,1));
}
}
// ----------------------------------------------------------------------------------------
void make_modularity_matrices (
const std::vector<sample_pair>& edges,
matrix<double>& A,
matrix<double>& P,
double& m
)
{
const unsigned long num_nodes = max_index_plus_one(edges);
A.set_size(num_nodes, num_nodes);
P.set_size(num_nodes, num_nodes);
A = 0;
P = 0;
std::vector<double> k(num_nodes,0);
for (unsigned long i = 0; i < edges.size(); ++i)
{
const unsigned long n1 = edges[i].index1();
const unsigned long n2 = edges[i].index2();
k[n1] += edges[i].distance();
if (n1 != n2)
{
k[n2] += edges[i].distance();
A(n2,n1) += edges[i].distance();
}
A(n1,n2) += edges[i].distance();
}
m = sum(A)/2;
for (long r = 0; r < P.nr(); ++r)
{
for (long c = 0; c < P.nc(); ++c)
{
P(r,c) = k[r]*k[c]/(2*m);
}
}
}
double compute_modularity_simple (
const std::vector<sample_pair>& edges,
std::vector<unsigned long> labels
)
{
double m;
matrix<double> A,P;
make_modularity_matrices(edges, A, P, m);
matrix<double> B = A - P;
double Q = 0;
for (long r = 0; r < B.nr(); ++r)
{
for (long c = 0; c < B.nc(); ++c)
{
if (labels[r] == labels[c])
{
Q += B(r,c);
}
}
}
return 1.0/(2*m) * Q;
}
// ----------------------------------------------------------------------------------------
void test_modularity(dlib::rand& rnd)
{
print_spinner();
std::vector<sample_pair> edges;
std::vector<ordered_sample_pair> oedges;
std::vector<unsigned long> labels;
make_test_graph(rnd, edges, labels, 10, 30, 3, 0.10);
if (rnd.get_random_double() < 0.5)
remove_duplicate_edges(edges);
convert_unordered_to_ordered(edges, oedges);
const double m1 = modularity(edges, labels);
const double m2 = compute_modularity_simple(edges, labels);
const double m3 = modularity(oedges, labels);
DLIB_TEST(std::abs(m1-m2) < 1e-12);
DLIB_TEST(std::abs(m2-m3) < 1e-12);
DLIB_TEST(std::abs(m3-m1) < 1e-12);
}
void test_newman_clustering(dlib::rand& rnd)
{
print_spinner();
std::vector<sample_pair> edges;
std::vector<unsigned long> labels;
make_test_graph(rnd, edges, labels, 5, 30, 3, 0.10);
if (rnd.get_random_double() < 0.5)
remove_duplicate_edges(edges);
std::vector<unsigned long> labels2;
unsigned long num_clusters = newman_cluster(edges, labels2);
DLIB_TEST(labels.size() == labels2.size());
DLIB_TEST(num_clusters == 5);
for (unsigned long i = 0; i < labels.size(); ++i)
{
for (unsigned long j = 0; j < labels.size(); ++j)
{
if (labels[i] == labels[j])
{
DLIB_TEST(labels2[i] == labels2[j]);
}
else
{
DLIB_TEST(labels2[i] != labels2[j]);
}
}
}
}
void test_chinese_whispers(dlib::rand& rnd)
{
print_spinner();
std::vector<sample_pair> edges;
std::vector<unsigned long> labels;
make_test_graph(rnd, edges, labels, 5, 30, 3, 0.10);
if (rnd.get_random_double() < 0.5)
remove_duplicate_edges(edges);
std::vector<unsigned long> labels2;
unsigned long num_clusters;
if (rnd.get_random_double() < 0.5)
num_clusters = chinese_whispers(edges, labels2, 200, rnd);
else
num_clusters = chinese_whispers(edges, labels2);
DLIB_TEST(labels.size() == labels2.size());
DLIB_TEST(num_clusters == 5);
for (unsigned long i = 0; i < labels.size(); ++i)
{
for (unsigned long j = 0; j < labels.size(); ++j)
{
if (labels[i] == labels[j])
{
DLIB_TEST(labels2[i] == labels2[j]);
}
else
{
DLIB_TEST(labels2[i] != labels2[j]);
}
}
}
}
void test_bottom_up_clustering()
{
std::vector<dpoint> pts;
pts.push_back(dpoint(0.0,0.0));
pts.push_back(dpoint(0.5,0.0));
pts.push_back(dpoint(0.5,0.5));
pts.push_back(dpoint(0.0,0.5));
pts.push_back(dpoint(3.0,3.0));
pts.push_back(dpoint(3.5,3.0));
pts.push_back(dpoint(3.5,3.5));
pts.push_back(dpoint(3.0,3.5));
pts.push_back(dpoint(7.0,7.0));
pts.push_back(dpoint(7.5,7.0));
pts.push_back(dpoint(7.5,7.5));
pts.push_back(dpoint(7.0,7.5));
matrix<double> dists(pts.size(), pts.size());
for (long r = 0; r < dists.nr(); ++r)
for (long c = 0; c < dists.nc(); ++c)
dists(r,c) = length(pts[r]-pts[c]);
matrix<unsigned long,0,1> truth(12);
truth = 0, 0, 0, 0,
1, 1, 1, 1,
2, 2, 2, 2;
std::vector<unsigned long> labels;
DLIB_TEST(bottom_up_cluster(dists, labels, 3) == 3);
DLIB_TEST(mat(labels) == truth);
DLIB_TEST(bottom_up_cluster(dists, labels, 1, 4.0) == 3);
DLIB_TEST(mat(labels) == truth);
DLIB_TEST(bottom_up_cluster(dists, labels, 1, 4.95) == 2);
truth = 0, 0, 0, 0,
0, 0, 0, 0,
1, 1, 1, 1;
DLIB_TEST(mat(labels) == truth);
DLIB_TEST(bottom_up_cluster(dists, labels, 1) == 1);
truth = 0, 0, 0, 0,
0, 0, 0, 0,
0, 0, 0, 0;
DLIB_TEST(mat(labels) == truth);
dists.set_size(0,0);
DLIB_TEST(bottom_up_cluster(dists, labels, 3) == 0);
DLIB_TEST(labels.size() == 0);
DLIB_TEST(bottom_up_cluster(dists, labels, 1) == 0);
DLIB_TEST(labels.size() == 0);
dists.set_size(1,1);
dists = 1;
DLIB_TEST(bottom_up_cluster(dists, labels, 3) == 1);
DLIB_TEST(labels.size() == 1);
DLIB_TEST(labels[0] == 0);
DLIB_TEST(bottom_up_cluster(dists, labels, 1) == 1);
DLIB_TEST(labels.size() == 1);
DLIB_TEST(labels[0] == 0);
DLIB_TEST(bottom_up_cluster(dists, labels, 1, 0) == 1);
DLIB_TEST(labels.size() == 1);
DLIB_TEST(labels[0] == 0);
dists.set_size(2,2);
dists = 1;
DLIB_TEST(bottom_up_cluster(dists, labels, 3) == 2);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(labels[0] == 0);
DLIB_TEST(labels[1] == 1);
DLIB_TEST(bottom_up_cluster(dists, labels, 1) == 1);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(labels[0] == 0);
DLIB_TEST(labels[1] == 0);
DLIB_TEST(bottom_up_cluster(dists, labels, 1, 1) == 1);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(labels[0] == 0);
DLIB_TEST(labels[1] == 0);
DLIB_TEST(bottom_up_cluster(dists, labels, 1, 0.999) == 2);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(labels[0] == 0);
DLIB_TEST(labels[1] == 1);
}
void test_segment_number_line()
{
dlib::rand rnd;
std::vector<double> x;
for (int i = 0; i < 5000; ++i)
{
x.push_back(rnd.get_double_in_range(-1.5, -1.01));
x.push_back(rnd.get_double_in_range(-0.99, -0.01));
x.push_back(rnd.get_double_in_range(0.01, 1));
}
auto r = segment_number_line(x,1);
std::sort(r.begin(), r.end());
DLIB_TEST(r.size() == 3);
DLIB_TEST(-1.5 <= r[0].lower && r[0].lower < r[0].upper && r[0].upper <= -1.01);
DLIB_TEST(-0.99 <= r[1].lower && r[1].lower < r[1].upper && r[1].upper <= -0.01);
DLIB_TEST(0.01 <= r[2].lower && r[2].lower < r[2].upper && r[2].upper <= 1);
x.clear();
for (int i = 0; i < 5000; ++i)
{
x.push_back(rnd.get_double_in_range(-2, 1));
x.push_back(rnd.get_double_in_range(-2, 1));
x.push_back(rnd.get_double_in_range(-2, 1));
}
r = segment_number_line(x,1);
DLIB_TEST(r.size() == 3);
r = segment_number_line(x,1.5);
DLIB_TEST(r.size() == 2);
r = segment_number_line(x,10.5);
DLIB_TEST(r.size() == 1);
DLIB_TEST(-2 <= r[0].lower && r[0].lower < r[0].upper && r[0].upper <= 1);
}
class test_clustering : public tester
{
public:
test_clustering (
) :
tester ("test_clustering",
"Runs tests on the clustering routines.")
{}
void perform_test (
)
{
test_bottom_up_clustering();
test_segment_number_line();
dlib::rand rnd;
std::vector<sample_pair> edges;
std::vector<unsigned long> labels;
DLIB_TEST(newman_cluster(edges, labels) == 0);
DLIB_TEST(chinese_whispers(edges, labels) == 0);
edges.push_back(sample_pair(0,1,1));
DLIB_TEST(newman_cluster(edges, labels) == 1);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(chinese_whispers(edges, labels) == 1);
DLIB_TEST(labels.size() == 2);
edges.clear();
edges.push_back(sample_pair(0,0,1));
DLIB_TEST(newman_cluster(edges, labels) == 1);
DLIB_TEST(labels.size() == 1);
DLIB_TEST(chinese_whispers(edges, labels) == 1);
DLIB_TEST(labels.size() == 1);
edges.clear();
edges.push_back(sample_pair(1,1,1));
DLIB_TEST(newman_cluster(edges, labels) == 1);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(chinese_whispers(edges, labels) == 2);
DLIB_TEST(labels.size() == 2);
edges.push_back(sample_pair(0,0,1));
DLIB_TEST(newman_cluster(edges, labels) == 2);
DLIB_TEST(labels.size() == 2);
DLIB_TEST(chinese_whispers(edges, labels) == 2);
DLIB_TEST(labels.size() == 2);
for (int i = 0; i < 10; ++i)
test_modularity(rnd);
for (int i = 0; i < 10; ++i)
test_newman_clustering(rnd);
for (int i = 0; i < 10; ++i)
test_chinese_whispers(rnd);
}
} a;
}