def test_predict(self):
"""
Tests the predict method using known model
params
"""
X, y, params, exp_y = self.solve_model()
# use cost function predict method
cost = cost_functions.GaussianCostFunction(X, y)
obs_y = cost.predict(X, params)
self.assertEqual(exp_y.shape, obs_y.shape)
def test_error(self):
"""
Tests the mean-squared error implementation
"""
X, y, _, pred_y = self.solve_model()
# use cost function predict method
cost = cost_functions.GaussianCostFunction(X, y)
obs_mse = cost._mse(y, pred_y)
exp_mse = mean_squared_error(y, pred_y)
self.assertAlmostEqual(obs_mse, exp_mse)
def test_cost(self):
"""
Tests the cost function.
"""
X, y, params, pred_y = self.solve_model()
exp_mse = mean_squared_error(y, pred_y)
# use cost function predict method
cost_func = cost_functions.GaussianCostFunction(X, y)
cost = cost_func.cost(params)
self.assertAlmostEqual(cost, exp_mse)
def test_optim(self):
"""
Test optimizing the cost function.
"""
X, y, exp_params, exp_pred_y = self.solve_model()
exp_mse = mean_squared_error(y, exp_pred_y)
cost_func = cost_functions.GaussianCostFunction(X, y)
optimizer = optim.Optimizer(0.1, 10000, 1e-4, 1e-4)
# use a mean of 0, variance of 1 to start
initial_params = np.array([0.0, 1.0])
optim_params, n_iters = optimizer.optimize(cost_func, initial_params)
print(n_iters)
print(exp_params)
print(optim_params)
self.assertAlmostEqual(optim_params[0], exp_params[0], 1)
self.assertAlmostEqual(optim_params[1], exp_params[1], 1)