def test_create_line_hyp_space():
n_features = 2
line_hyp_space_one = create_line_hyp_space(n_features)
true_hyp_space_one = np.array([[1, 0], [0, 1], [1, 1]])
assert np.array_equal(line_hyp_space_one, true_hyp_space_one)
n_features = 4
line_hyp_space_two = create_line_hyp_space(n_features)
true_hyp_space_two = np.array([[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0],
[0, 0, 0, 1], [1, 1, 0, 0], [0, 1, 1, 0],
[0, 0, 1, 1], [1, 1, 1, 0], [0, 1, 1, 1],
[1, 1, 1, 1]])
assert np.array_equal(line_hyp_space_two, true_hyp_space_two)
def __init__(self, n_features=3, hyp_space_type="boundary",
sampling="max", true_hyp=None):
assert(n_features > 0)
self.d = [] # observed data points
self.n_obs = 0 # number of observed data points
self.n_labels = 2 # number of possible y values
self.n_features = n_features
if hyp_space_type == "boundary":
self.hyp_space = utils.create_boundary_hyp_space(self.n_features)
elif hyp_space_type == "line":
self.hyp_space = utils.create_line_hyp_space(self.n_features)
self.n_hyp = len(self.hyp_space)
self.prior = np.array([1 / self.n_hyp
for _ in range(self.n_hyp)])
self.posterior = self.prior
if true_hyp is not None:
self.true_hyp = true_hyp
self.true_hyp_idx = \
np.where([np.all(true_hyp == hyp)
for hyp in self.hyp_space])[0]
else:
self.true_hyp_idx = np.random.randint(self.n_hyp)
self.true_hyp = self.hyp_space[self.true_hyp_idx]
self.posterior_true_hyp = np.ones(self.n_features + 1)
self.posterior_true_hyp[0] = 1 / self.n_hyp
self.first_feature_prob = np.zeros(n_features)
self.sampling = sampling
def __init__(self, n_features=3, hyp_space_type="boundary",
sampling="max", true_hyp=None):
self.n_features = n_features
self.n_labels = 2
self.observed_features = np.array([])
self.observed_labels = np.array([])
self.n_obs = 0
self.features = np.arange(self.n_features)
self.labels = np.arange(self.n_labels)
if hyp_space_type == "boundary":
self.hyp_space = utils.create_boundary_hyp_space(self.n_features)
elif hyp_space_type == "line":
self.hyp_space = utils.create_line_hyp_space(self.n_features)
self.n_hyp = len(self.hyp_space)
self.learner_prior = (1 / self.n_hyp) * \
np.ones((self.n_hyp, self.n_features, self.n_labels))
self.self_teaching_posterior = np.zeros(
(self.n_hyp, self.n_features, self.n_labels))
self.learner_posterior = self.learner_prior
self.sampling = sampling
if true_hyp is not None:
self.true_hyp = true_hyp
self.true_hyp_idx = \
np.where([np.all(true_hyp == hyp)
for hyp in self.hyp_space])[0]
else:
self.true_hyp_idx = np.random.randint(self.n_hyp)
self.true_hyp = self.hyp_space[self.true_hyp_idx]
self.posterior_true_hyp = np.ones(self.n_features + 1)
self.posterior_true_hyp[0] = 1 / self.n_hyp
self.first_feature_prob = np.zeros(self.n_features)
def __init__(self, n_features=3, hyp_space_type="boundary"):
assert(n_features > 0)
self.n_features = n_features
self.n_labels = 2 # number of possible y values
if hyp_space_type == "boundary":
self.hyp_space = utils.create_boundary_hyp_space(self.n_features)
elif hyp_space_type == "line":
self.hyp_space = utils.create_line_hyp_space(self.n_features)
self.n_hyp = len(self.hyp_space)
self.prior = 1 / self.n_hyp * \
np.ones((self.n_hyp, self.n_features, self.n_labels))
def __init__(self, n_features=3, hyp_space_type="boundary"):
if hyp_space_type == "boundary":
self.hyp_space = utils.create_boundary_hyp_space(n_features)
elif hyp_space_type == "line":
self.hyp_space = utils.create_line_hyp_space(n_features)
self.n_features = n_features
self.n_labels = 2
self.features = np.arange(self.n_features)
self.labels = np.arange(self.n_labels)
self.n_hyp = len(self.hyp_space)
self.learner_prior = (1 / self.n_hyp) * \
np.ones((self.n_hyp, self.n_features, self.n_labels))
self.self_teaching_posterior = np.zeros(
(self.n_hyp, self.n_features, self.n_labels))
self.learner_posterior = self.learner_prior