def test_Precision(self):
data = Table('iris')
learner = LogisticRegressionLearner()
results = TestOnTrainingData(data, [learner])
self.assertGreater(Precision(results)[0], 0.9)
self.assertEqual(round(Precision(results)[0], 3), 0.928)
def setUpClass(cls):
cls.iris = Table('iris')
cls.score = Precision()
def test_Precision(self):
data = Table('iris')
learner = LogisticRegressionLearner(preprocessors=[])
results = TestOnTrainingData(data, [learner])
self.assertAlmostEqual(Precision(results)[0], 0.962, 3)
# then train the model, and produce results.
# So, simply initialize the CrossValidation() object from the 'testing' library
# and call it with input arguments 1) the dataset and 2) the learner.
# Note that the 'learner' argument should be in array form, i.e. '[learner]'.
cv = CrossValidation(k=10, random_state=44)
results = {}
print('Training')
for name, learner in tqdm(learners.items()):
results[name] = cv(data=wineData, learners=[learner])
# As for the required metrics, you can get them using the 'evaluation.scoring' library.
# The 'average' parameter of each metric is used while measuring scores to perform
# a type of averaging on the data. DON'T WORRY MUCH ABOUT THAT JUST YET (AGAIN). USE EITHER
# 'MICRO' OR 'MACRO' (preferably 'macro', at least for final results).
# =============================================================================
# # ADD COMMANDS TO EVALUATE YOUR MODEL HERE (AND PRINT ON CONSOLE)
scores = defaultdict(dict)
print('Testing')
for name, res in tqdm(results.items()):
scores[name]['acc'] = CA(results=res)[0]
scores[name]['prec'] = Precision(results=res, average='macro')[0]
scores[name]['rec'] = Recall(results=res, average='macro')[0]
scores[name]['f1'] = F1(results=res, average='macro')[0]
classifier = learners[name](data=wineData)
rules = '\n'.join([str(rule) for rule in classifier.rule_list])
scores[name]['rules'] = rules
scores_df = pd.DataFrame.from_dict(data=scores).transpose()
scores_df.to_csv(category + '_' + evaluator + '.csv')