AHCClassifier Class Reference

#include <ahc_classifier.h>

Collaboration diagram for AHCClassifier:

Public Member Functions
void	clusterData (const Matrix2D< double > &X, int numberOfClusters=2, int distance=2, int linkageType=1)
void	clusterWithDistance (const Matrix2D< double > &D, int numberOfClusters=2, int linkageType=1)

Public Attributes
Matrix1D< int >	clusterAssigned
std::vector< std::vector< int > >	cluster

Detailed Description

AHC classifier class.

Definition at line 38 of file ahc_classifier.h.

Member Function Documentation

clusterData()

void AHCClassifier::clusterData	(	const Matrix2D< double > &	X,
		int	numberOfClusters = 2,
		int	distance = 2,
		int	linkageType = 1
	)

Cluster data. X is the data to classify, each row is an observation. Columns are features of that observation.

For the distances see http://www.alglib.net/translator/man/manual.cpp.html#sub_clusterizersetpoints 0 Chebyshev distance (L-inf norm) 1 city block distance (L1 norm) 2 Euclidean distance (L2 norm) 10 Pearson correlation: dist(a,b) = 1-corr(a,b) 11 Absolute Pearson correlation: dist(a,b) = 1-|corr(a,b)| 12 Uncentered Pearson correlation (cosine of the angle): dist(a,b) = a'b/(|a||b|) 13 Absolute uncentered Pearson correlation dist(a,b) = |a'b|/(|a||b|) 20 Spearman rank correlation: dist(a,b) = 1-rankcorr(a,b) 21 Absolute Spearman rank correlation dist(a,b) = 1-|rankcorr(a,b)|

For linkage: 0 complete linkage (default algorithm) 1 single linkage 2 unweighted average linkage 3 weighted average linkage

Definition at line 31 of file ahc_classifier.cpp.

{
    alglib::integer_1d_array cidx;
    try
    {
        // See example at http://www.alglib.net/translator/man/manual.cpp.html#example_clst_ahc
        alglib::real_2d_array xy;
        xy.setcontent(MAT_YSIZE(X),MAT_XSIZE(X),MATRIX2D_ARRAY(X));

        alglib::clusterizerstate s;
        alglib::ahcreport rep;

        // Run Hierarchical Clustering
        clusterizercreate(s);
        clusterizersetahcalgo(s, linkageType);
        clusterizersetpoints(s, xy, distance);
        clusterizerrunahc(s, rep);

        // Now get K clusters
        alglib::integer_1d_array cz;
        clusterizergetkclusters(rep, numberOfClusters, cidx, cz);
    }
    catch (alglib::ap_error e)
    {
        REPORT_ERROR(ERR_UNCLASSIFIED,e.msg);
    }

    // Collect results
    clusterAssigned.resizeNoCopy(MAT_YSIZE(X));
    FOR_ALL_ELEMENTS_IN_MATRIX1D(clusterAssigned)
        VEC_ELEM(clusterAssigned,i)=cidx[i];

    std::vector<int> dummy;
    cluster.clear();
    for (int k=0; k<numberOfClusters; ++k)
        cluster.push_back(dummy);
    FOR_ALL_ELEMENTS_IN_MATRIX1D(clusterAssigned)
        cluster[VEC_ELEM(clusterAssigned,i)].push_back(i);
}

clusterWithDistance()

void AHCClassifier::clusterWithDistance	(	const Matrix2D< double > &	D,
		int	numberOfClusters = 2,
		int	linkageType = 1
	)

Cluster given a distance matrix. X is the data to classify, each row is an observation. Columns are features of that observation.

For linkage: 0 complete linkage (default algorithm) 1 single linkage 2 unweighted average linkage 3 weighted average linkage

Definition at line 71 of file ahc_classifier.cpp.

{
    alglib::integer_1d_array cidx;
    try
    {
        // See example at https://www.tol-project.org/svn/tolp/OfficialTolArchiveNetwork/AlgLib/CppTools/source/alglib/manual.cpp.html#example_clst_kclusters
        alglib::real_2d_array d;
        d.setcontent(MAT_YSIZE(D),MAT_XSIZE(D),MATRIX2D_ARRAY(D));

        alglib::clusterizerstate s;
        alglib::ahcreport rep;

        // Run Hierarchical Clustering
        clusterizercreate(s);
        clusterizersetahcalgo(s, linkageType);
        clusterizersetdistances(s, d, true);
        clusterizerrunahc(s, rep);

        // Now get K clusters
        alglib::integer_1d_array cz;
        clusterizergetkclusters(rep, numberOfClusters, cidx, cz);
    }
    catch (alglib::ap_error e)
    {
        REPORT_ERROR(ERR_UNCLASSIFIED,e.msg);
    }

    // Collect results
    clusterAssigned.resizeNoCopy(MAT_YSIZE(D));
    FOR_ALL_ELEMENTS_IN_MATRIX1D(clusterAssigned)
        VEC_ELEM(clusterAssigned,i)=cidx[i];

    std::vector<int> dummy;
    cluster.clear();
    for (int k=0; k<numberOfClusters; ++k)
        cluster.push_back(dummy);
    FOR_ALL_ELEMENTS_IN_MATRIX1D(clusterAssigned)
        cluster[VEC_ELEM(clusterAssigned,i)].push_back(i);
}

Member Data Documentation

cluster

std::vector< std::vector<int> > AHCClassifier::cluster

There are as many elements in the vector as clusters. Each vector in the cluster contains the list of individuals assigned to this cluster.

Definition at line 49 of file ahc_classifier.h.

clusterAssigned

Matrix1D<int> AHCClassifier::clusterAssigned

The i-th location is the number of the cluster of the i-th element in the input data. The length of this vector is the number of individuals.

Definition at line 44 of file ahc_classifier.h.

---

The documentation for this class was generated from the following files:

• xmipp/libraries/classification/ahc_classifier.h
• xmipp/libraries/classification/ahc_classifier.cpp