def test_sample_choice(self):
prob = array([[0.25, 0.25, 0.25, 0.25],
[0, 0, 1, 0],
[0.99, 0, 0.001, 0.009]])
start_time = time.time()
sample = sample_choice(prob)
logger.log_status("sample_choice from (%s, %s) items array in " % prob.shape + str(time.time() - start_time) + " sec")
self.assertEqual(sample[0].size, prob.shape[0], msg ="sample size not equal to sample size parameter")
self.assertEqual(prob[sample].size, prob.shape[0], msg ="sample size not equal to sample size parameter")
assert 0 <= sample[1].min() <= prob.shape[1]-1, "sampled elements not in between min and max of source array"
assert 0 <= sample[1].max() <= prob.shape[1]-1, "sampled elements not in between min and max of source array"
assert all(not_equal(prob[sample], 0.0)), "elements with zero weight in the sample"
def run(self, probability, resources=None):
""" Compute choices according to given probability.
'probability' is a 2D numpy array (nobservations x nequations).
The returned value is a 1D array of choice indices [0, nequations-1] of length nobservations.
"""
if probability.ndim == 1: # if 1d array, add column of 1-probability
probability = reshape(probability, (probability.size,1))
probability = concatenate((1.0-probability, probability), axis=1)
if probability.ndim < 2:
raise StandardError, "Argument 'probability' must be a 2D numpy array."
prob = probability/reshape(probability.sum(axis=1),(probability.shape[0],1))
return sample_choice(prob)[1]
def run(self, probability, resources=None):
""" Compute choices according to given probability.
'probability' is a 2D numpy array (nobservations x nequations).
The returned value is a 1D array of choice indices [0, nequations-1] of length nobservations.
"""
if probability.ndim == 1: # if 1d array, add column of 1-probability
probability = reshape(probability, (probability.size, 1))
probability = concatenate((1.0 - probability, probability), axis=1)
if probability.ndim < 2:
raise StandardError, "Argument 'probability' must be a 2D numpy array."
prob = probability / reshape(probability.sum(axis=1),
(probability.shape[0], 1))
return sample_choice(prob)[1]
