Package rdkit :: Package ML :: Package InfoTheory :: Module BitClusterer :: Class BitClusterer
[hide private]
[frames] | no frames]

type BitClusterer

source code

object --+
         |
        BitClusterer

Class to cluster a set of bits based on their correllation

The correlation matrix is first built using by reading the fingerprints
from a database or a list of fingerprints

Instance Methods [hide private]
 
__init__(self, idList, nCluster, type=rdkit.SimDivFilters.rdSimDivPickers.ClusterMethod.WARD) source code
 
ClusterBits(self, corrMat) source code
 
SetClusters(self, clusters) source code
 
GetClusters(self) source code
 
MapToClusterScores(self, fp)
Map the fingerprint to a real valued vector of score based on the bit clusters
source code
 
MapToClusterFP(self, fp)
Map the fingerprint to a smaller sized (= number of clusters) fingerprint
source code
Method Details [hide private]

__init__(self, idList, nCluster, type=rdkit.SimDivFilters.rdSimDivPickers.ClusterMethod.WARD)
(Constructor)

source code 
Overrides: object.__init__
(inherited documentation)

MapToClusterScores(self, fp)

source code 
Map the fingerprint to a real valued vector of score based on the bit clusters

The dimension of the vector is same as the number of clusters. Each value in the 
vector corresponds to the number of bits in the corresponding cluster
that are turned on in the fingerprint

ARGUMENTS:
 - fp : the fingerprint 

MapToClusterFP(self, fp)

source code 
Map the fingerprint to a smaller sized (= number of clusters) fingerprint

Each cluster get a bit in the new fingerprint and is turned on if any of the bits in
the cluster are turned on in the original fingerprint