langmore / statsmodels Goto Github PK

After pull #5 was merged, 3 tests don't pass. There are differences at around the 4th decimal place. I am using numpy 1.6.1 and scipy 0.9.0.

-------------------- Test result snippets ----------------------------------

FAIL: statsmodels.discrete.tests.test_discrete.TestPoissonL1Compatability.test_f_test
...
AssertionError:
Arrays are not almost equal to 3 decimals

(mismatch 100.0%)
x: array([[ 11794.18165332]])
y: array([[ 11794.18460073]])

FAIL: statsmodels.discrete.tests.test_discrete.TestProbitCG.test_conf_int
...
AssertionError:
Arrays are not almost equal to 4 decimals

(mismatch 37.5%)
x: array([[ 0.26591211, 2.98595073],
[ -0.11269725, 0.2161498 ],
[ 0.26008461, 2.59262668],
[-12.43592369, -2.46939961]])
y: array([[ 0.2658255, 2.985795 ],
[ -0.1126929, 0.2161508],
[ 0.2600795, 2.592585 ],
[-12.43547 , -2.469166 ]])

FAIL: statsmodels.discrete.tests.test_discrete.TestProbitCG.test_params
...
AssertionError:
Arrays are not almost equal to 4 decimals

(mismatch 25.0%)
x: array([ 1.62593142, 0.05172627, 1.42635565, -7.45266165])
y: array([ 1.62581025, 0.05172895, 1.42633237, -7.45232042])

cov_params() does not work with L1CountResults (result returned by sm.Poisson.fit_regularized. It raises the following:

TypeError: unbound method cov_params() must be called with CountResults instance as first argument (got L1CountResults instance instead)

Using zero start params may cause issues. In any case, different values should be explored.

if start_params is None:
start_params = np.zeros((self.K * (self.J-1)))
else:
start_params = np.asarray(start_params)