def testFitDiscretePowerLaw2(self):
try:
import networkx
except ImportError:
logging.debug("Networkx not found, can't run test")
return
nxGraph = networkx.barabasi_albert_graph(1000, 2)
graph = SparseGraph.fromNetworkXGraph(nxGraph)
degreeSeq = graph.outDegreeSequence()
output = Util.fitDiscretePowerLaw(degreeSeq)
def testFitDiscretePowerLaw(self):
#Test with small x
x = numpy.array([5])
ks, alpha2, xmin = Util.fitDiscretePowerLaw(x)
self.assertEquals(ks, -1)
self.assertEquals(alpha2, -1)
x = numpy.array([5, 2])
ks, alpha2, xmin = Util.fitDiscretePowerLaw(x)
#Test with a large vector x
alpha = 2.5
exponent = (1/(alpha-1))
numPoints = 15000
x = 10*numpy.random.rand(numPoints)**-exponent
x = numpy.array(numpy.round(x), numpy.int)
x = x[x<=500]
x = x[x>=1]
xmins = numpy.arange(1, 15)
ks, alpha2, xmin = Util.fitDiscretePowerLaw(x, xmins)
self.assertAlmostEqual(alpha, alpha2, places=1)
def fitPowerLaw(self):
"""
Fits the out-degree probabilities of this graph using the power law
p_d ~ d^-gamma. The value of xmin is the point to start taking examples.
:returns alpha: The power law exponent.
:returns ks: A fit of the power law curve to the data using KS.
:returns xmin: The minimum value of x.
"""
if self.getNumEdges() == 0:
return 0,0,0
logging.warn("Use with caution: fitPowerLaw may not be stable")
degreeSeq = self.outDegreeSequence()
degreeMax = numpy.max(degreeSeq)
xmins = numpy.arange(1, numpy.minimum(20, degreeMax))
ks, alpha, xmin = Util.fitDiscretePowerLaw(degreeSeq, xmins)
return alpha, ks, xmin
