Module mixture :: Class ConstrainedDataSet
[hide private]
[frames] | no frames]

Class ConstrainedDataSet

source code


Extension of the DataSet object that can hold pairwise or label constraints in the objects. This data set is required for the semi-supervised learning EM

Instance Methods [hide private]
 
__init__(self)
Creates and returns an empty DataSet object
source code
 
__copy__(self)
Interface to copy.copy function.
source code
 
setConstrainedLabels(self, labels)
Sets labels for semi-supervised learning.
source code
 
setPairwiseConstraints(self, positive, negative)
Set pairwise constraints.
source code

Inherited from DataSet: __len__, __str__, extractSubset, filterSamples, fromArray, fromFiles, fromList, getExternalFeature, getInternalFeature, getMissingIndices, internalInit, maskDataSet, maskFeatures, printClustering, removeFeatures, removeSamples, setMissingSymbols, singleFeatureSubset, writeClusteringFasta

Method Details [hide private]

__init__(self)
(Constructor)

source code 

Creates and returns an empty DataSet object

Overrides: DataSet.__init__
(inherited documentation)

__copy__(self)

source code 

Interface to copy.copy function.

Returns:
deep copy of 'self'
Overrides: DataSet.__copy__
(inherited documentation)

setConstrainedLabels(self, labels)

source code 

Sets labels for semi-supervised learning.

Parameters:
  • labels - list of lists of sample indices. The index in 'labels' denotes the component the samples are assigned to. For instance labels = [[0,2],[4,6]] would mean samples 0 and 2 are labelled with component 0.

setPairwiseConstraints(self, positive, negative)

source code 

Set pairwise constraints.

XXX add params