Commit 595c545a authored by Joe Grimes's avatar Joe Grimes
Browse files

Updated python script, multi class testing for sil scoring

parent 432d31ab

src/loss_function/LossFunctionSilhouetteScore.cpp

+4 −2
Original line number Diff line number Diff line
@@ -88,6 +88,7 @@ LossFunctionSilhouetteScore::LossFunctionSilhouetteScore(const std::shared_ptr<L 


void LossFunctionSilhouetteScore::set_nfeat(int n_feat)

{

    _silhouette_scores.clear();

    _n_feat = n_feat;

    _a.resize(_n_feat * _n_samp);

    _n_dim = n_feat + 1;
@@ -145,7 +146,7 @@ double LossFunctionSilhouetteScore::calculate_a_i(size_t samp_idx, int cluster_l 
    return sum / static_cast<double>(n_samples_in_cluster - 1);

}



// b_i for sample samp_idx: min average distance to other clusters



double LossFunctionSilhouetteScore::calculate_b_i(size_t sample_idx, int cluster_label) {

    const double* p1 = &_a[sample_idx * _n_feat];

    double best_avg = std::numeric_limits<double>::infinity();
@@ -174,7 +175,7 @@ void LossFunctionSilhouetteScore::populate_sil_score() { 
    for (auto& [cluster_label, indices] : _cluster_indices) {

        // step by _n_feat to visit each sample once

        for (size_t i = 0; i < indices.size(); i += _n_feat) {

            size_t sample_idx = indices[i] / _n_feat;  // recover row/sample index

            size_t sample_idx = indices[i] / _n_feat;

            a_i = calculate_a_i(sample_idx, cluster_label);

            b_i = calculate_b_i(sample_idx, cluster_label);

            double s_i = 0.0;
@@ -210,6 +211,7 @@ double LossFunctionSilhouetteScore::test_loss(const std::vector<model_node_ptr>& 
}



double LossFunctionSilhouetteScore::get_silhouette_average() {

    std::cout << std::endl;

    if (_silhouette_scores.empty()) return 0.0;

    return std::accumulate(_silhouette_scores.begin(), _silhouette_scores.end(), 0.0)

           / static_cast<double>(_silhouette_scores.size());

src/loss_function/silhouette.py

+21 −6
Original line number Diff line number Diff line
@@ -60,22 +60,37 @@ def run_feature_testing(train_feat1, train_feat2, cluster_labels): 




if __name__ == "__main__":

    #Double cluster tests:

    labels = np.array([0, 0, 1, 1])

    if True:

    if False:

        #For feats ~0.466666666

        train_feat1 = np.array([4.0, 5.0, 6.0, 7.0])

        train_feat2 = np.array([2.0, 3.0, 4.0, 5.0])



        run_feature_testing(train_feat1, train_feat2, labels)



    if True:

    if False:

        #values ripped from d_matrix

        dat = np.array([

            [-1.86846, -1.9923, -1.95254, -1.56359],

            [-1.78104, -1.31323, -1.24359, -1.72509],

            [-1.06531, -1.47307, -1.01745, -1.10234],

            [-1.96543, -1.29881, -1.92731, -1.49548]

            [-1.86846, -1.9923],

            [-1.78104, -1.31323],

            [-1.06531, -1.47307],

            [-1.96543, -1.29881]

        ])



        score = silhouette_score(dat, labels)

        print("Silhouette score for inds:", score)

    if True:

        X = np.array([

            [1.0, 2.0], [1.2, 2.1], [0.8, 1.9],

            [5.0, 5.0], [5.1, 5.2], [4.9, 4.8],

            [9.0, 1.0], [9.1, 1.1], [8.9, 0.9],

            [2.0, 8.0], [2.1, 8.2], [1.9, 7.8]

        ])



        # Cluster labels

        labels = np.array([0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3])



        # Compute silhouette score

        sil_score = silhouette_score(X, labels)

        print("Silhouette Score:", sil_score)
 
 No newline at end of file

tests/googletest/loss_function/test_silhouette_score_loss.cc

+34 −0
Original line number Diff line number Diff line
@@ -197,6 +197,40 @@ TEST_F(LossFunctionSilhouetteScoreTests, ManualDeterminationFeats) 
    EXPECT_NEAR(silhouette_score,0.466666, 0.0001);

}



TEST_F(LossFunctionSilhouetteScoreTests, MultiClassTest)

{

    _task_sizes_train = {12};

    _task_sizes_test = {12};

    node_value_arrs::finalize_values_arr();

    node_value_arrs::initialize_values_arr(_task_sizes_train, _task_sizes_test, 2, 2, false);



    std::vector<double> train_feat1 = {1.0,1.2,0.8, 5.0,5.1,4.9, 9.0,9.1,8.9, 2.0,2.1,1.9};

    std::vector<double> train_feat2 = {2.0,2.1,1.9, 5.0,5.2,4.8, 1.0,1.1,0.9, 8.0,8.2,7.8};



    std::vector<double> test_feat1 = train_feat1;

    std::vector<double> test_feat2 = train_feat2;



    _phi.clear();

    _model_phi.clear();



    _phi.push_back(std::make_shared<FeatureNode>(0, "Feature1", train_feat1, test_feat1, Unit("m")));

    _phi.push_back(std::make_shared<FeatureNode>(1, "Feature2", train_feat2, test_feat2, Unit("m")));



    _model_phi.push_back(std::make_shared<ModelNode>(_phi[0]));

    _model_phi.push_back(std::make_shared<ModelNode>(_phi[1]));



    // 4 classes

    _prop_train = {0,0,0, 1,1,1, 2,2,2, 3,3,3};

    _prop_test = _prop_train;



    LossFunctionSilhouetteScore loss(

        _prop_train, _prop_test, _task_sizes_train, _task_sizes_test, false, 2);



    double silhouette_score = loss(_model_phi);



    EXPECT_NEAR(silhouette_score, 0.941, 0.001);

}



}