class SVMTest(unittest.TestCase):
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/lin_training.csv")
self.problem.set_label("Name")
self.problem.set_model("SVM")
self.data = self.problem.data
self.model = self.problem.model
def tearDown(self):
self.problem = None
self.data = None
def test_fit_stats(self):
self.model.fit(None)
acc = calculate_acc(self.model.fit_report['output'], self.model.fit_report['target'])
self.assertEqual(0.97668, round(acc, 5))
class ModelUtilTest(unittest.TestCase):
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/iris.csv")
self.problem.set_label("Name")
self.problem.set_model("NaiveBayes")
self.data = self.problem.data
def tearDown(self):
self.problem = None
self.data = None
def test_unique_list(self):
ulist = unique_list(self.data['Name'])
self.assertEqual(set(['Iris-versicolor', 'Iris-setosa']), set(ulist))
def test_get_both_columns(self):
num_cols, cat_cols = get_both_columns(self.data, 'Name')
self.assertEqual(set(['SepalLength', 'SepalWidth', 'PetalLength', 'PetalWidth']), set(num_cols))
self.assertEqual(set(['Dum']), set(cat_cols))
def test_get_mean_std(self):
mean, std = get_mean_std(self.data['SepalWidth'])
self.assertEqual(3.0940, round(mean, 4))
self.assertEqual(0.4761, round(std, 4))
def test_calculate_prob(self):
prob = calculate_prob(10, 2, 3)
self.assertEqual(0.0038, round(prob, 4))
def test_nonlin(self):
out = nonlin(0.5, False)
self.assertEqual(0.62246, round(out, 5))
out = nonlin(0.5, True)
self.assertEqual(0.25, round(out, 5))
def test_get_class(self):
x = get_class(0.3)
self.assertEqual(0, x)
def test_get_class_np(self):
x = get_class_np(-0.3)
self.assertEqual(-1, x)
class NeuralNetworkTest(unittest.TestCase):
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/iris.csv")
self.problem.set_label("Name")
self.problem.set_model("NeuralNetwork")
self.data = self.problem.data
self.model = self.problem.model
def tearDown(self):
self.problem = None
self.data = None
def test_fit_stats(self):
self.model.fit()
acc = calculate_acc(self.model.fit_report['output'], self.model.fit_report['target'])
self.assertEqual(1.0, acc)
def test_fit_stats_oneIter(self):
self.model.num_iter = 1
self.model.fit()
acc = calculate_acc(self.model.fit_report['output'], self.model.fit_report['target'])
self.assertEqual(.5, acc)
class NaiveBayesTest(unittest.TestCase):
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/iris.csv")
self.problem.set_label("Name")
self.problem.set_model("NaiveBayes")
self.data = self.problem.data
self.model = self.problem.model
def tearDown(self):
self.problem = None
self.data = None
def test_prob_stats(self):
prob = self.model.prob_hub['Dum']['Iris-versicolor']['a']
self.assertEqual(0.40, round(prob, 2))
prob = self.model.prob_hub['Dum']['Iris-setosa']['b']
self.assertEqual(0.26, round(prob, 2))
def test_mean_std(self):
mean, std = get_mean_std(self.data['SepalWidth'])
self.assertEqual(3.0940, round(mean, 4))
self.assertEqual(0.4761, round(std, 4))
class ProblemTest(unittest.TestCase):
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/iris.csv")
self.problem.set_label("Name")
def tearDown(self):
self.problem = None
def test_data_read(self):
self.assertEqual(len(self.problem.data), 100)
def test_problem_reload(self):
self.problem.save("test_problem.json", "test_data.json")
self.problem.load("test_problem.json", "test_data.json")
self.test_problem_descriptions()
def test_problem_descriptions(self):
self.assertEqual(self.problem.label, "Name")
self.assertEqual(self.problem.problem_type, "BinaryClassification")
self.assertEqual(self.problem.file_path, "../../../examples/data/iris.csv")
__author__ = 'Jiarui Xu'
from learnpy.Problem import Problem
# create a problem with training data
pro = Problem("BinaryClassification", "./data/iris_training.csv")
# set the predictor variable
pro.set_label('Name')
# save the problem
pro.save("problem.json", "data.json")
# load the problem
pro.load("problem.json", "data.json")
pro.set_model("SVM")
pro.model.fit(None)
# set testing data
pro.set_testing("./data/iris_testing.csv")
pro.predict()
# new problem for 242 demo
# create a problem
pro2 = Problem("BinaryClassification", "./data/lin_training.csv")
# set the predictor variable
pro2.set_label('Name')
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/iris.csv")
self.problem.set_label("Name")
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/iris.csv")
self.problem.set_label("Name")
self.problem.set_model("NeuralNetwork")
self.data = self.problem.data
self.model = self.problem.model
__author__ = 'Jiarui Xu'
from learnpy.Problem import Problem
# create a problem
pro = Problem("BinaryClassification", "./data/iris_training.csv")
# set the predictor variable
pro.set_label('Name')
pro.set_model("NaiveBayes")
pro.model.fit(None)
pro.set_testing("./data/iris_testing.csv")
pro.predict()
# new problem
# create a problem
pro2 = Problem("BinaryClassification", "./data/lin_training.csv")
# set the predictor variable
pro2.set_label('Name')
pro2.set_model("NaiveBayes")
pro2.model.fit(None)
pro2.set_testing("./data/lin_testing.csv")
pro2.predict()
__author__ = 'Jiarui Xu'
from learnpy.Problem import Problem
# create a problem
pro = Problem("BinaryClassification", "./data/iris_training.csv")
# set the predictor variable
pro.set_label('Name')
# use the model neural network
pro.set_model("NeuralNetwork", num_iter=1000)
# fit the model into the data set above
pro.model.fit(None)
# set testing data
pro.set_testing("./data/iris_testing.csv")
pro.predict()
def setUp(self):
self.problem = Problem("BinaryClassification", "../../../examples/data/lin_training.csv")
self.problem.set_label("Name")
self.problem.set_model("SVM")
self.data = self.problem.data
self.model = self.problem.model