RecursivePartitionalClusteringMetric.h

Go to the documentation of this file.

/**
 *  \file IMP/statistics/RecursivePartitionalClusteringMetric.h
 *  \brief Metric for clustering data that has already been clustered once.
 *
 *  Copyright 2007-2016 IMP Inventors. All rights reserved.
 *
 */

#ifndef IMPSTATISTICS_RECURSIVE_PARTITIONAL_CLUSTERING_METRIC_H
#define IMPSTATISTICS_RECURSIVE_PARTITIONAL_CLUSTERING_METRIC_H

#include <IMP/statistics/statistics_config.h>
#include "Metric.h"
#include "PartitionalClustering.h"
#include <IMP/Pointer.h>

IMPSTATISTICS_BEGIN_NAMESPACE

//! Represent a metric for clustering data that has already been clustered once.
/** To use it, cluster things once, then create one of these with the metric
    you want (created with the original data). When you pass this metric to
    the clustering algorithm, it will cluster the centers. You can extract the
    clustering of the original elements using create_full_clustering().
*/
class IMPSTATISTICSEXPORT RecursivePartitionalClusteringMetric : public Metric {
  IMP::PointerMember<Metric> metric_;
  IMP::PointerMember<PartitionalClustering> clustering_;

 public:
  RecursivePartitionalClusteringMetric(Metric *metric,
                                       PartitionalClustering *clustering);
  PartitionalClustering *create_full_clustering(
      PartitionalClustering *center_cluster);
  double get_distance(unsigned int i, unsigned int j) const IMP_OVERRIDE;
  unsigned int get_number_of_items() const IMP_OVERRIDE;
  IMP_OBJECT_METHODS(RecursivePartitionalClusteringMetric);
};

IMPSTATISTICS_END_NAMESPACE

#endif /* IMPSTATISTICS_RECURSIVE_PARTITIONAL_CLUSTERING_METRIC_H */