1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
|
//
// Copyright (C) 2003-2006 Rational Discovery LLC
//
// @@ All Rights Reserved @@
// This file is part of the RDKit.
// The contents are covered by the terms of the BSD license
// which is included in the file license.txt, found at the root
// of the RDKit source tree.
//
#include <RDGeneral/export.h>
#ifndef _HIERARCHCLUSTERPICKER_H
#define _HIERARCHCLUSTERPICKER_H
#include <RDGeneral/types.h>
#include "DistPicker.h"
namespace RDPickers {
/*! \brief Diversity picker based on hierarchical clustering
*
* This class inherits from DistPicker since it uses the distance matrix
* for diversity picking. The clustering itself is done using the Murtagh
* code in $RDBASE/Code/ML/Cluster/Mutagh/
*/
class RDKIT_SIMDIVPICKERS_EXPORT HierarchicalClusterPicker : public DistPicker {
public:
/*! \brief The type of hierarchical clustering algorithm to use
*/
typedef enum {
WARD = 1,
SLINK = 2,
CLINK = 3,
UPGMA = 4,
MCQUITTY = 5,
GOWER = 6,
CENTROID = 7
} ClusterMethod;
/*! \brief Constructor - takes a ClusterMethod as an argument
*
* Sets the hierarch clustering method
*/
explicit HierarchicalClusterPicker(ClusterMethod clusterMethod)
: d_method(clusterMethod) {
;
};
/*! \brief This is the function that does the picking
*
* Here is how the algorithm works \n
* FIX: Supply reference
*
* - The entire pool is clustered using the distance matrix using one of the
* hierachical clustering method (specified via the constructor). \n
* - Starting with the individaul items in the pool, clusters are merged based
* on the output from clustering method. \n
* - The merging is stopped when the number of clusters is same as
* the number of picks.
* - For each item in a cluster the sum of square of the distances to the rest
*of
* of the items (in the cluster) is computed. The item with the smallest of
*values is
* picked as a representative of the cluster. Basically trying to pick the
*item closest
* to the centroid of the cluster.
*
*
* \param distMat - distance matrix - a vector of double. It is assumed
*that only the
* lower triangle element of the matrix are supplied in a 1D
*array\n
* NOTE: this matrix WILL BE ALTERED during the picking\n
* \param poolSize - the size of the pool to pick the items from. It is
*assumed that the
* distance matrix above contains the right number of elements;
*i.e.
* poolSize*(poolSize-1) \n
* \param pickSize - the number items to pick from pool (<= poolSize)
*/
RDKit::INT_VECT pick(const double *distMat, unsigned int poolSize,
unsigned int pickSize) const;
/*! \brief This is the function that does the clustering of the items - used
*by the picker
*
* ARGUMENTS:
*
* \param distMat - distance matrix - a vector of double. It is assumed that
*only the
* lower triangle element of the matrix are supplied in a 1D
*array\n
* NOTE: this matrix WILL BE ALTERED during the picking\n
* \param poolSize - the size of the pool to pick the items from. It is
*assumed that the
* distance matrix above contains the right number of elements;
*i.e.
* poolSize*(poolSize-1) \n
* \param pickSize - the number clusters to divide the pool into (<=
*poolSize)
*/
RDKit::VECT_INT_VECT cluster(const double *distMat, unsigned int poolSize,
unsigned int pickSize) const;
private:
ClusterMethod d_method;
};
};
#endif
|
