def get_models(args):
settings = []
if args.small is not None:
# Small
for weight in ['uniform', 'distance']:
for k in range(1, 6, 1):
setting = {
'n_neighbors': k,
'weights': weight,
}
settings.append(setting)
else:
# Large
for weight in ['uniform', 'distance']:
for k in range(1, 21, 1):
setting = {
'n_neighbors': k,
'weights': weight,
}
settings.append(setting)
settings = cls.override_settings(args, settings, KNeighborsClassifier)
models = cls.models_from_settings(settings, KNeighborsClassifier)
return models
def get_models(args):
settings = []
if args.small is not None:
# Small
setting = {
'alpha': 3**-4,
'hidden_layer_sizes': (800, 200, 30),
'random_state': 1,
'activation': 'tanh',
'max_iter': 10000
}
settings.append(setting)
else:
setting = {
'alpha': 3**-4,
'hidden_layer_sizes': (800, 200, 30),
'random_state': 1,
'activation': 'tanh',
'max_iter': 10000
}
settings.append(setting)
settings = cls.override_settings(args, settings, MLPClassifier)
models = cls.models_from_settings(settings, MLPClassifier)
return models
def get_models(args):
settings = []
if args.small is not None:
# Small
setting = {}
settings.append(setting)
else:
# Large
setting = {}
settings.append(setting)
settings = cls.override_settings(args, settings,
QuadraticDiscriminantAnalysis)
models = cls.models_from_settings(settings, QuadraticDiscriminantAnalysis)
return models
def get_models(args):
settings = []
if args.small is not None:
# Small
for C in range(-5, 6):
setting = {'max_iter': 1000, 'C': 3**C}
settings.append(setting)
else:
# Large
for C in range(-5, 6):
setting = {'max_iter': 1000, 'C': 3**C}
settings.append(setting)
settings = cls.override_settings(args, settings, LogisticRegression)
models = cls.models_from_settings(settings, LogisticRegression)
return models
def get_models(args):
settings = []
if args.small is not None:
# Small
for gamma in range(-4, -2):
for C in range(-5, 6):
setting = {'kernel': 'rbf', 'gamma': 10**gamma, 'C': 3**C}
settings.append(setting)
else:
# Large
for gamma in range(-4, -2):
for C in range(-3, -1):
setting = {'kernel': 'rbf', 'gamma': 10**gamma, 'C': 3**C}
settings.append(setting)
settings = cls.override_settings(args, settings, SVC)
models = cls.models_from_settings(settings, SVC)
return models