def test_two_filters(self):
""" Tests saturation_filter (normal and prune) and classification_filter
"""
import cf_noise_detection.library as l
import cf_weka.classification as c
learner = c.j48()
data = ut.load_UCI_dataset("iris")
inp_dict = {'data': data, 'satur_type': 'normal'}
out_dict = l.saturation_filter(inp_dict, None)
self.assertGreaterEqual(len(out_dict.keys()), 1)
inp_dict = {'data': data, 'satur_type': 'prune'}
out_dict = l.saturation_filter(inp_dict, None)
self.assertGreaterEqual(len(out_dict.keys()), 1)
inp_dict = {
'learner': learner,
'data': data,
'timeout': 60.0,
'k_folds': 10
}
out_dict = l.classification_filter(inp_dict, None)
self.assertGreaterEqual(len(out_dict.keys()), 1)
def test_two_filters(self):
""" Tests saturation_filter (normal and prune) and classification_filter
"""
import cf_noise_detection.library as l
import cf_weka.classification as c
learner = c.j48()
data = ut.load_UCI_dataset("iris")
inp_dict = {'data': data, 'satur_type': 'normal'}
out_dict = l.saturation_filter(inp_dict, None)
self.assertGreaterEqual(len(out_dict.keys()), 1)
inp_dict = {'data': data, 'satur_type': 'prune'}
out_dict = l.saturation_filter(inp_dict, None)
self.assertGreaterEqual(len(out_dict.keys()), 1)
inp_dict = {'learner': learner, 'data': data, 'timeout': 60.0, 'k_folds': 10}
out_dict = l.classification_filter(inp_dict, None)
self.assertGreaterEqual(len(out_dict.keys()), 1)
def test_noise_rank(self):
"""Tests noise rank widget
"""
import cf_noise_detection.library as l
import cf_weka.classification as c
learner = c.j48()
data = ut.load_UCI_dataset("iris")
inp_dict = {'learner': learner, 'data': data, 'timeout': 60.0, 'k_folds': 10}
out_dict = l.classification_filter(inp_dict, None)
inp_dict = {'noise': [out_dict['noise_dict']],
'data': data}
out_dict = l.noiserank(inp_dict)
self.assertGreaterEqual(len(out_dict['allnoise']), 1)
def test_noise_rank(self):
"""Tests noise rank widget
"""
import cf_noise_detection.library as l
import cf_weka.classification as c
learner = c.j48()
data = ut.load_UCI_dataset("iris")
inp_dict = {
'learner': learner,
'data': data,
'timeout': 60.0,
'k_folds': 10
}
out_dict = l.classification_filter(inp_dict, None)
inp_dict = {'noise': [out_dict['noise_dict']], 'data': data}
out_dict = l.noiserank(inp_dict)
self.assertGreaterEqual(len(out_dict['allnoise']), 1)
def test_classification_learners(self):
""" Tests creating classification learners from the classification.py file
:return: a list of learners, i.e. WekaClassifier objects
"""
lrn = []
num_exceptions = 0
try:
lrn.append(c.logistic())
lrn.append(c.j48())
lrn.append(c.rep_tree())
lrn.append(c.random_forest())
lrn.append(c.random_tree())
lrn.append(c.rules_zeror())
lrn.append(c.rules_jrip())
lrn.append(c.ibk())
lrn.append(c.k_star())
lrn.append(c.naive_bayes())
lrn.append(c.multilayer_perceptron())
lrn.append(c.smo())
except Exception, e:
num_exceptions += 1
print "Exception: " + str(e)