A Data Instance stores the points it is generated from a DataDescriptor
classNumber
(int
): The number of classes in the DataInstance
nPoints
(int
): The number of points in the DataInstance
points
([DataPoint]
): The points of the DataInstance
dimension
(int
): The dimension of the DataInstance
. Corresponds to the dimension
of the DataPoint
in points
##Constructors
__init__({'classNumber': classNumber, 'nPoints': nPoints, 'points': points, 'dimension': dimension, 'orientation': orientation})
plot(noBack=False, nPoints=self.nPoints):
Plot nPoints
randomly selected of the DataInstance
. If noBack
, the points with class 0
are ignord
numpyify():
return numpy.array([numpy.array, numpy.array])
Returns numpy arrays of the points and labels split
predict(model):
return DataInstance
Returns the DataInstance
predicted by the Keras + Tensorflow network when feeding the instance points
predictAndEvaluate(model):
return float
Computes the accuracy of the Keras + Tensorflow network over the DataInstance
computeBottleNeckDistance(DataInstance, min_persistence=0, nPoints=self.nPoints, targetCluster=[1]):
return float
Computes the bottleneckDistance of the DataInstance
to another one passed in argument using up to nPoints
from the instances and restricting the study to points with classes in targetCluster
bettiNumbers(targetCluster=[1], threshold=0.05, nPoints=self.nPoints, maxDim=self.dimension, maxEdge=10, fromValue=0.05, toValue=0.05):
return [int]
Computes the betti numbers of the DataInstance
using gudhi
for more information go to gudhi's documentation
newBettiNumbers(targetCluster=[1], threshold=0.05, nPoints=self.nPoints, errorRate=0.005, plot=False, showProgress=False):
return [int]
Computes the betti numbers of the DataInstance
using a custom algorithm for up to nPoints
evenly distributed in targetCluster