def model_complexity_exp(self):
#TODO should we create a new learner object??
self.learner = DTLearner()
self.expHelper = ExperimentHelper(self.splitter, self.learner)
param_range = np.array([2, 4, 6, 8])
self.expHelper.model_complexity_exp('max_depth', param_range)
self.learner = DTLearner(max_depth=4)
self.expHelper = ExperimentHelper(self.splitter, self.learner)
param_range = np.array([2, 5, 6, 8, 10])
self.expHelper.model_complexity_exp('max_leaf_nodes', param_range)
"""self.learner = DTLearner(max_depth = 4, max_leaf_nodes = 5)
def model_complexity_exp2(self):
#TODO should we create a new learner object??
self.learner = SVMLearner(kernel='linear', C=0.2)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'linear4', 'Linear Kernel')
param_range = np.array([1, 2, 4, 5, 6, 7, 8])
self.expHelper.model_complexity_exp('gamma', param_range)
self.learner = SVMLearner(kernel='linear', C=0.2)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'linear5', 'Linear Kernel')
param_range = np.array([0.1, 0.2, 0.4, 0.5, 0.6, 0.7, 0.8])
self.expHelper.model_complexity_exp('gamma', param_range)
def model_complexity_exp_alpha1(self):
#TODO should we create a new learner object??
self.learner = ANNLearner(hidden_layer_sizes=(300, ))
self.expHelper = ExperimentHelper(self.splitter, self.learner, '1')
#param_range = np.array([(100),(200,),(300,),(400,),(500,)])
#param_range = np.array([0.0001, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1])
param_range = np.array([0.0001, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1])
self.expHelper.model_complexity_exp('alpha', param_range)
self.learner = ANNLearner(hidden_layer_sizes=(390, ))
self.expHelper = ExperimentHelper(self.splitter, self.learner, '2')
#param_range = np.array([(100),(200,),(300,),(400,),(500,)])
param_range = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7])
#param_range = np.array([0.0001, 0.0005, 0.001, 0.005, 0.008])
self.expHelper.model_complexity_exp('alpha', param_range)
def model_complexity_exp_2(self):
#TODO should we create a new learner object??
self.learner = BoostLearner(max_depth=2)
self.expHelper = ExperimentHelper(self.splitter, self.learner)
param_range = np.array([0.1, 0.2, 0.4, 0.6, 0.8, 0.9])
self.expHelper.model_complexity_exp('learning_rate', param_range)
def model_complexity_exp(self):
#TODO should we create a new learner object??
self.learner = BoostLearner()
self.expHelper = ExperimentHelper(self.splitter, self.learner)
param_range = np.array([2, 4, 6, 8, 9, 10])
self.expHelper.model_complexity_exp('max_depth', param_range)
def model_complexity_exp_linear1(self):
#TODO should we create a new learner object??
self.learner = LinearSVMLearner(dual=True)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'linear-2', 'LinearKernel')
param_range = np.array([0.01, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8])
self.expHelper.model_complexity_exp('C', param_range)
def model_complexity_exp_linear2(self):
#TODO should we create a new learner object??
self.learner = LinearSVMLearner(fit_intercept=True, C=0.01)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'linear-3', 'LinearKernel')
param_range = np.array([1, 2, 3, 4, 5, 6])
self.expHelper.model_complexity_exp('intercept_scaling', param_range)
def model_complexity_exp_alpha2(self):
#TODO should we create a new learner object??
self.learner = ANNLearner()
self.expHelper = ExperimentHelper(self.splitter, self.learner, '2')
#param_range = np.array([(100),(200,),(300,),(400,),(500,)])
param_range = np.array([0.0001, 0.001, 0.01, 0.1, 1, 5])
self.expHelper.model_complexity_exp('alpha', param_range)
def learning_curve_iter2_bank(self):
self.learner = KNNLearner(metric='euclidean',
n_neighbors=38,
weights='uniform')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'-iter-2')
self.expHelper.learning_curve_exp()
def model_complexity_exp3(self):
#TODO should we create a new learner object??
self.learner = KNNLearner(weights='distance')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'distance weight')
param_range = np.array([1, 2, 3, 4, 5])
self.expHelper.model_complexity_exp('n_neighbors', param_range)
def model_complexity_exp11(self):
#TODO should we create a new learner object??
self.learner = ANNLearner()
self.expHelper = ExperimentHelper(self.splitter, self.learner)
param_range = np.array([(50, ), (
50,
50,
), (
50,
50,
50,
), (
50,
50,
50,
50,
), (
50,
50,
50,
50,
50,
)])
#param_range = np.array([100, 200,300,400, 500])
self.expHelper.model_complexity_exp('hidden_layer_sizes', param_range)
def model_complexity_exp2(self):
#TODO should we create a new learner object??
self.learner = KNNLearner(metric='euclidean')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'euclidean')
param_range = np.array([1, 2, 3, 4, 5])
self.expHelper.model_complexity_exp('n_neighbors', param_range)
def experiment_run_test_bank_linear(self):
self.learner = LinearSVMLearner(fit_intercept=True,
C=0.01,
intercept_scaling=6)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'Linear Kernel')
self.expHelper.experiment_run_test()
def model_complexity_exp4(self):
#TODO should we create a new learner object??
self.learner = KNNLearner(metric='euclidean', weights='distance')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'distance weight')
param_range = np.arange(1, 40, 2)
print(param_range)
self.expHelper.model_complexity_exp('n_neighbors', param_range)
def learning_curve_iter2_bank_linear(self):
print("linear")
self.learner = LinearSVMLearner(fit_intercept=True,
C=0.01,
intercept_scaling=6)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'Linear Kernel')
self.expHelper.learning_curve_exp()
def model_complexity_exp11(self):
#TODO should we create a new learner object??
self.learner = KNNLearner()
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'euclidean', '1')
#param_range = np.array([1,2,3,4,5])
param_range = np.arange(40, 60, 2)
print(param_range)
self.expHelper.model_complexity_exp('n_neighbors', param_range)
def experiment_run_test_bank2(self):
self.learner = ANNLearner(activation='relu',
alpha=0.7,
hidden_layer_sizes=(390, ),
learning_rate='constant',
solver='adam',
early_stopping=False)
self.expHelper = ExperimentHelper(self.splitter, self.learner)
self.expHelper.experiment_run_test()
def model_complexity_exp4(self):
#TODO should we create a new learner object??
self.learner = SVMLearner(kernel='linear')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'linear3', 'Linear Kernel')
#param_range = np.array([0.5, 1, 1.5, 2])
#param_range = np.array([2, 3, 4, 5, 6, 9, 12])
#param_range = np.array([0.01, 0.02, 0.03, 0.043, 0.05, 0.06])
param_range = np.array([0.0001, 0.0005, 0.0008, 0.001, 0.005, 0.008])
self.expHelper.model_complexity_exp('C', param_range)
def learning_curve_iter21_bank(self):
self.learner = ANNLearner(activation='relu',
alpha=0.01,
hidden_layer_sizes=(300, ),
learning_rate='constant',
solver='adam',
early_stopping=False)
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'-iter-2')
self.expHelper.learning_curve_exp()
def model_complexity_exp44(self):
#TODO should we create a new learner object??
self.learner = SVMLearner(kernel='linear')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'linear1', 'Linear Kernel')
#param_range = np.array([0.5, 1, 1.5, 2])
#param_range = np.array([2, 3, 4, 5, 6, 9, 12])
param_range = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.7, 0.8, 0.9, 1])
self.expHelper.model_complexity_exp('C', param_range)
"""self.learner = SVMLearner(kernel = 'linear')
def model_complexity_exp_epoch2(self):
#TODO should we create a new learner object??
self.learner = ANNLearner(hidden_layer_sizes=(300, ),
alpha=0.0001,
early_stopping=False)
self.expHelper = ExperimentHelper(self.splitter, self.learner, '3')
#param_range = np.array([(100),(200,),(300,),(400,),(500,)])
#param_range = np.array([0.0001, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1])
param_range = np.array([1, 10, 50, 100, 200, 500, 1000])
self.expHelper.model_complexity_exp('max_iter', param_range)
def model_complexity_exp3(self):
#TODO should we create a new learner object??
self.learner = SVMLearner(kernel='sigmoid')
self.expHelper = ExperimentHelper(self.splitter, self.learner,
'sigmoid', 'Sigmoid Kernel')
#param_range = np.array([0.05, 0.1, 0.2, 0.3, 0.4, 0.5])
#self.expHelper.model_complexity_exp('C', param_range)
param_range = np.array([0.01, 0.03, 0.05, 0.07, 0.09])
self.expHelper.model_complexity_exp('gamma', param_range)
def model_complexity_exp6(self):
#TODO should we create a new learner object??
#self.learner = SVMLearner(kernel = 'rbf', C =0.1)
self.learner = SVMLearner(kernel='rbf', C=0.5)
self.expHelper = ExperimentHelper(self.splitter, self.learner, 'rbf8',
'rbf Kernel')
#param_range = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7])
#param_range = np.array([0.5,2, 3, 4, 5, 6])
param_range = np.array([0.01, 0.03, 0.05, 0.07, 0.09])
self.expHelper.model_complexity_exp('gamma', param_range)
self.learner = SVMLearner(kernel='rbf', C=0.5)
self.expHelper = ExperimentHelper(self.splitter, self.learner, 'rbf9',
'rbf Kernel')
param_range = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7])
#param_range = np.array([0.5,2, 3, 4, 5, 6])
#param_range = np.array([0.01, 0.03, 0.05, 0.07, 0.09])
self.expHelper.model_complexity_exp('gamma', param_range)
"""self.learner = SVMLearner(kernel = 'rbf', C =0.2)
def model_complexity_exp(self):
scoring = 'accuracy'
nn_model1 = mlrose.NeuralNetwork(hidden_nodes = [50,50,50,50], activation = 'relu', \
algorithm = 'random_hill_climb', max_iters = 1000, \
bias = True, is_classifier = True, learning_rate = 0.0001, \
early_stopping = True, clip_max = 5, max_attempts = 100, \
random_state = 3)
expHelper = ExperimentHelper(self.splitter, nn_model1, 'RHC')
param_range = np.array([0.0001, 0.001, 0.002, 0.003, 0.005, 0.008])
#param_range = np.array([100, 200,300,400, 500])
expHelper.model_complexity_exp('learning_rate', param_range)
def model_complexity_exp_restart_rhc(self):
#TODO should we create a new learner object??
scoring = 'accuracy'
nn_model1 = mlrose.NeuralNetwork(hidden_nodes = [50,50,50,50], activation = 'relu', \
algorithm = 'random_hill_climb', max_iters = 1000, \
bias = True, is_classifier = True, learning_rate = 0.0001, \
early_stopping = True, clip_max = 5, max_attempts = 100, \
random_state = 3, restarts = 0)
expHelper = ExperimentHelper(self.splitter, nn_model1, 'RHC')
param_range = [1, 10, 50, 100, 200, 300, 400, 500]
expHelper.model_complexity_exp('restarts', param_range)
def model_complexity_exp_max_attempt_ga(self):
#TODO should we create a new learner object??
scoring = 'accuracy'
nn_model1 = mlrose.NeuralNetwork(hidden_nodes = [50,50,50,50], activation = 'relu', \
algorithm = 'genetic_alg', max_iters = 1000, \
bias = True, is_classifier = True, learning_rate = 0.0001, \
early_stopping = True, clip_max = 5, max_attempts = 100, \
pop_size= 2000, random_state = 3)
expHelper = ExperimentHelper(self.splitter, nn_model1, 'GA')
param_range = [20, 50, 100, 200, 300]
expHelper.model_complexity_exp('max_attempts', param_range)
def model_complexity_exp_epoch_no_stop_sa(self):
#TODO should we create a new learner object??
scoring = 'accuracy'
nn_model1 = mlrose.NeuralNetwork(hidden_nodes = [50,50,50,50], activation = 'relu', \
algorithm = 'simulated_annealing', max_iters = 1000, \
bias = True, is_classifier = True, learning_rate = 0.0001, \
early_stopping = False, clip_max = 5, max_attempts = 100, \
random_state = 3)
expHelper = ExperimentHelper(self.splitter, nn_model1, 'SA')
param_range = [1, 10, 50, 100, 200, 500]
expHelper.model_complexity_exp('max_iters', param_range)
def experiment_run_test_bank_iter2(self):
self.learner = ANNLearner(activation='relu',
alpha=0.06,
hidden_layer_sizes=(200, 200, 200, 200, 200,
200, 200),
learning_rate='constant',
solver='adam',
early_stopping=True,
max_iter=600,
momentum=0.4)
print('100, 100, 100, 100, 100, 100,50 alpha 0.3')
self.expHelper = ExperimentHelper(self.splitter, self.learner)
self.expHelper.experiment_run_test()
""" self.learner = ANNLearner(
def model_complexity_exp5(self):
#TODO should we create a new learner object??
self.learner = SVMLearner(kernel='rbf')
self.expHelper = ExperimentHelper(self.splitter, self.learner, 'rbf1',
'rbf Kernel')
param_range = np.array([0.05, 0.1, 0.2, 0.3, 0.4, 0.5])
self.expHelper.model_complexity_exp('C', param_range)
self.learner = SVMLearner(kernel='rbf')
self.expHelper = ExperimentHelper(self.splitter, self.learner, 'rbf2',
'rbf Kernel')
param_range = np.array([0.5, 1, 1.5, 2])
self.expHelper.model_complexity_exp('C', param_range)
self.learner = SVMLearner(kernel='rbf')
self.expHelper = ExperimentHelper(self.splitter, self.learner, 'rbf3',
'rbf Kernel')
param_range = np.array([2, 3, 4, 5, 6, 9, 12])
self.expHelper.model_complexity_exp('C', param_range)
def model_complexity_exp_max_attempts_sa(self):
#TODO should we create a new learner object??
scoring = 'accuracy'
decay = mlrose.decay.ExpDecay()
nn_model1 = mlrose.NeuralNetwork(hidden_nodes = [50,50,50,50], activation = 'relu', \
algorithm = 'simulated_annealing', max_iters = 1000, \
bias = True, is_classifier = True, learning_rate = 0.0001, \
early_stopping = True, clip_max = 5, max_attempts = 100, \
schedule = decay,
random_state = 3)
expHelper = ExperimentHelper(self.splitter, nn_model1, 'SA')
param_range = [20, 50, 100, 200, 300]
expHelper.model_complexity_exp('max_attempts', param_range)
print('completed max attempt sa nn')