def __init__(self, parameters={}):
self.params = utils.update_dictionary_items(
{
'regwgt': 0.01,
'features': range(385),
'tolerance': 10e-4,
}, parameters)
def __init__(self, parameters = {}):
# Default parameters, any of which can be overwritten by values passed to params
#self.params = utils.update_dictionary_items({'regwgt': 0.5}, parameters)
self.params = utils.update_dictionary_items({
'regwgt': 0.5, # l2 regularizer
'features': [1,2,3,4,5],
}, parameters)
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items(
{
'stepsize': 0.01,
'epochs': 1000
}, parameters)
self.weights = None
def __init__(self, parameters={}):
# Default parameters, any of which can be overwritten by values passed to params
self.params = utils.update_dictionary_items(
{
'regwgt': 0.01,
'features': range(385),
}, parameters)
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items(
{
'regwgt': 0.01,
'stepsize': 0.01,
'tolerance': 10e-4,
'maxiteration': 1000
}, parameters)
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items(
{
'regwgt': 0.0,
'features': [1, 2, 3, 4, 5],
#'features': list(range(1, 385)),
},
parameters)
def __init__(self, parameters = {}, iterations = 1000000, tolerance = 10e-4):
self.params = utils.update_dictionary_items({
'iterations': 1000000,
'tolerance': 0.001,
}, parameters)
self.iterations = iterations
self.tolerance = tolerance
def __init__(self, parameters = {},step_size = 0.01, epochs = 1000):
self.params = utils.update_dictionary_items({
'step_size': 0.01,
'epochs': 1000,
}, parameters)
self.step_size = step_size
self.epochs = epochs
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items(
{
'stepsize': 0.01,
'epochs': 100,
'centers': 10,
'kernel': 'linear'
}, parameters)
self.weights = None
def __init__(self,parameters = {}, learning_rate = 0.001, iterations = 1000, lamda = 0.0005, tolerance = 10e-4):
self.params = utils.update_dictionary_items({
'learning_rate': 0.001,
'lamda': 0.0005,
}, parameters)
self.learning_rate = learning_rate
self.iterations = iterations
self.lamda = lamda
self.tolerance = tolerance
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items(
{
'regwgt': 0.01,
'features': range(385),
"epochs": 1000,
"stepsize": 0.01,
}, parameters)
self.noofruns = 5
self.error = np.zeros(1000)
def __init__(self, parameters={}):
# Default parameters, any of which can be overwritten by values passed to params
self.params = utils.update_dictionary_items(
{
'regwgt': 0.01,
'features': [1, 2, 3, 4, 5],
'step_size': 0.01
}, parameters)
self.weights = None
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items({
'nh': 4,
'transfer': 'sigmoid',
'stepsize': 0.01,
'epochs': 10,
}, parameters)
if self.params['transfer'] is 'sigmoid':
self.transfer = utils.sigmoid
self.dtransfer = utils.dsigmoid
else:
# For now, only allowing sigmoid transfer
raise Exception('NeuralNet -> can only handle sigmoid transfer, must set option transfer to string sigmoid')
self.wi = None
self.wo = None
def __init__(self, parameters={}):
# Default parameters, any of which can be overwritten by values passed to params
self.params = utils.update_dictionary_items({'regwgt': 0.5},
parameters)
self.weights = None
def __init__(self, parameters = {}):
self.params = utils.update_dictionary_items({
'regwgt': 0.5, # l2 regularizer
'features': [1,2,3,4,5],
}, parameters)
def __init__(self, parameters={}):
self.params = utils.update_dictionary_items({
"iteration": 1000,
}, parameters)
def __init__(self, parameters={}):
# Default parameters, any of which can be overwritten by values passed to params
self.params = utils.update_dictionary_items({'tau': 0.7}, parameters)
def __init__(self, parameters={}):
""" Params can contain any useful parameters for the algorithm """
# Assumes that a bias unit has been added to feature vector as the last feature
# If usecolumnones is False, it ignores this last feature
self.params = utils.update_dictionary_items({'usecolumnones': False},
parameters)
