def estimate_beta_ridge(x, y, alpha):
""" SGD estimate of beta by minimizing the squared error and the
ridge penalty scale alpha. """
beta_initial = [decimal.Decimal(random.random()) for _ in x[0]]
return minimize_stochastic(
partial(squared_error_ridge, alpha=alpha),
partial(squared_error_ridge_gradient, alpha=alpha), x, y, beta_initial,
decimal.Decimal(0.001))
def estimate_beta_ridge(x, y, alpha):
""" SGD estimate of beta by minimizing the squared error and the
ridge penalty scale alpha. """
beta_initial = [decimal.Decimal(random.random()) for _ in x[0]]
return minimize_stochastic(partial(squared_error_ridge, alpha=alpha),
partial(squared_error_ridge_gradient,
alpha=alpha), x, y, beta_initial,
decimal.Decimal(0.001))
def estimate_beta(x, y):
# guess the initial beta
#beta_initial = [decimal.Decimal(random.random()) for _ in x[0]]
#return minimize_stochastic(squared_error, squared_error_gradient,
# x, y, beta_initial, decimal.Decimal(0.001))
# guess the initial beta
beta_initial = [random.random() for _ in x[0]]
return minimize_stochastic(squared_error, squared_error_gradient, x, y,
beta_initial, 0.001)
def estimate_beta(x, y):
# guess the initial beta
#beta_initial = [decimal.Decimal(random.random()) for _ in x[0]]
#return minimize_stochastic(squared_error, squared_error_gradient,
# x, y, beta_initial, decimal.Decimal(0.001))
# guess the initial beta
beta_initial = [random.random() for _ in x[0]]
return minimize_stochastic(squared_error, squared_error_gradient,
x, y, beta_initial, 0.001)