def setUp(self):
self.data = data.fromfile(testfile('testdata10.txt'))
self.neteval = evaluator.NetworkEvaluator(
self.data,
network.fromdata(self.data),
prior.UniformPrior(self.data.variables.size))
self.neteval.network.edges.add_many([(1,0),(2,0),(3,0)])
def setUp(self):
a,b,c,d,e = 0,1,2,3,4
self.data = data.fromfile(testfile('testdata9.txt'))
self.net = network.fromdata(self.data)
self.net.edges.add_many([(a,c), (b,c), (c,d), (c,e)])
self.neteval1 = self.neteval_type(self.data, self.net, max_iterations="10*n**2")
def fromconfig(data_=None, network_=None, prior_=None):
"""Create an evaluator based on configuration parameters.
This function will return the correct evaluator based on the relevant
configuration parameters.
"""
data_ = data_ or data.fromconfig()
network_ = network_ or network.fromdata(data_)
prior_ = prior_ or prior.fromconfig()
if data_.missing.any():
e = _missingdata_evaluators[config.get('evaluator.missingdata_evaluator')]
return e(data_, network_, prior_)
else:
return SmartNetworkEvaluator(data_, network_, prior_)
def test_gibbs_scoring(self):
# score two nets (correct one and bad one) with missing values.
# ensure that correct one scores better. (can't check for exact score)
# score network: {a,b}->c->{d,e}
score1 = self.neteval1.score_network()
# score network: {a,b}->{d,e} c
a,b,c,d,e = 0,1,2,3,4
net2 = network.fromdata(self.data)
net2.edges.add_many([(a,d), (a,e), (b,d), (b,e)])
neteval2 = self.neteval_type(self.data, net2)
score2 = neteval2.score_network()
# score1 should be better than score2
print score1, score2
assert score1 > score2, "Gibbs sampling can find goodhidden node."
def setUp(self):
self.data = data.fromfile(testfile('testdata10.txt'))
self.ne = evaluator.SmartNetworkEvaluator(
self.data,
network.fromdata(self.data))
def setUp(self):
self.data = data.fromfile(testfile('testdata10.txt'))
self.neteval = evaluator.NetworkEvaluator(self.data, network.fromdata(self.data))
self.neteval.network.edges.add_many([(1,0), (2,0), (3,0)]) # {1,2,3} --> 0
