def setUp(self):
session.init("test_least_squared_error_loss")
self.lse_loss = LeastSquaredErrorLoss()
self.y_list = [i % 2 for i in range(100)]
self.predict_list = [random.random() for i in range(100)]
self.y = session.parallelize(self.y_list, include_key=False)
self.predict = session.parallelize(self.predict_list, include_key=False)
class TestLeastSquaredErrorLoss(unittest.TestCase):
def setUp(self):
session.init("test_least_squared_error_loss")
self.lse_loss = LeastSquaredErrorLoss()
self.y_list = [i % 2 for i in range(100)]
self.predict_list = [random.random() for i in range(100)]
self.y = session.parallelize(self.y_list, include_key=False)
self.predict = session.parallelize(self.predict_list, include_key=False)
def test_predict(self):
for y in self.y_list:
y_pred = self.lse_loss.predict(y)
self.assertTrue(np.fabs(y_pred - y) < consts.FLOAT_ZERO)
def test_compute_gradient(self):
for y, y_pred in zip(self.y_list, self.predict_list):
lse_grad = self.lse_loss.compute_grad(y, y_pred)
grad = 2 * (y_pred - y)
self.assertTrue(np.fabs(lse_grad - grad) < consts.FLOAT_ZERO)
def test_compute_hess(self):
for y, y_pred in zip(self.y_list, self.predict_list):
hess = 2
lse_hess = self.lse_loss.compute_hess(y, y_pred)
self.assertTrue(np.fabs(lse_hess - hess) < consts.FLOAT_ZERO)
def test_compute_loss(self):
sklearn_loss = metrics.mean_squared_error(self.y_list, self.predict_list)
lse_loss = self.lse_loss.compute_loss(self.y, self.predict)
self.assertTrue(np.fabs(lse_loss - sklearn_loss) < consts.FLOAT_ZERO)
def set_loss(self, objective_param):
loss_type = objective_param.objective
params = objective_param.params
LOGGER.info("set objective, objective is {}".format(loss_type))
if self.task_type == consts.CLASSIFICATION:
if loss_type == "cross_entropy":
if self.num_classes == 2:
self.loss = SigmoidBinaryCrossEntropyLoss()
else:
self.loss = SoftmaxCrossEntropyLoss()
else:
raise NotImplementedError("objective %s not supported yet" %
(loss_type))
elif self.task_type == consts.REGRESSION:
if loss_type == "lse":
self.loss = LeastSquaredErrorLoss()
elif loss_type == "lae":
self.loss = LeastAbsoluteErrorLoss()
elif loss_type == "huber":
self.loss = HuberLoss(params[0])
elif loss_type == "fair":
self.loss = FairLoss(params[0])
elif loss_type == "tweedie":
self.loss = TweedieLoss(params[0])
elif loss_type == "log_cosh":
self.loss = LogCoshLoss()
else:
raise NotImplementedError("objective %s not supported yet" %
(loss_type))
else:
raise NotImplementedError("objective %s not supported yet" %
(loss_type))