def setUp(self):
self.cyy = np.array([[1.0 ,0.554 ,0.227 ,0.189 ,0.461 ,0.506 ,0.408, 0.280 ,0.241],
[0.554 ,1.0 ,0.296 ,0.219 ,0.479 ,0.530 ,0.425, 0.311 ,0.311],
[0.277 ,0.296 ,1.0 ,0.769 ,0.237 ,0.243 ,0.304, 0.718 ,0.730],
[0.189 ,0.219 ,0.769 ,1.0 ,0.212 ,0.226 ,0.291, 0.681 ,0.661],
[0.461 ,0.479 ,0.237 ,0.212 ,1.0 ,0.520 ,0.514, 0.313 ,0.245],
[0.506 ,0.530 ,0.243 ,0.226 ,0.520 ,1.0 ,0.473, 0.348 ,0.290],
[0.408 ,0.425 ,0.304 ,0.291 ,0.514 ,0.473 ,1.0 , 0.374 ,0.306],
[0.280 ,0.311 ,0.718 ,0.681 ,0.313 ,0.348 ,0.374, 1.0 ,0.672],
[0.241 ,0.311 ,0.730 ,0.661 ,0.245 ,0.290 ,0.306, 0.672 ,1.0 ]])
self.lambdas = np.array([[0.70, 0.74, 0.39, 0.37, 0.65, 0.72, 0.60, 0.51, 0.48],
[-0.12,-0.08,0.81, 0.75,-0.03,-0.05, 0.09, 0.65, 0.67],
[0.15, 0.22, 0.33, 0.08, 0.00, 0.00, 0.00, 0.00, 0.00],
[0.00, 0.00, 0.00, 0.00, 0.37, 0.15, 0.35, 0.02,-0.12]])
self.taus = np.diag([0.48, 0.41, 0.09, 0.31, 0.44, 0.46, 0.52, 0.32, 0.32])
self.params_ex = {'lambdas' : self.lambdas, 'taus' : self.taus}
self.emmlfa_ex = EMMlFactorAnalysis(4,100,self.params_ex)
pattern = np.array([[1,1,1,1,1,1,1,1,1],
[1,1,1,1,1,1,1,1,1],
[1,1,1,1,0,0,0,0,0],
[0,0,0,0,1,1,1,1,1]])
self.params_cm = { 'lambdas' : self.lambdas, 'taus' : self.taus, 'factor_pattern' : pattern }
self.emmlfa_cm = EMMlFactorAnalysis(4,100,self.params_cm, fa_type="confirmatory")
class TestEMMlfa:
def setUp(self):
self.cyy = np.array([[1.0 ,0.554 ,0.227 ,0.189 ,0.461 ,0.506 ,0.408, 0.280 ,0.241],
[0.554 ,1.0 ,0.296 ,0.219 ,0.479 ,0.530 ,0.425, 0.311 ,0.311],
[0.277 ,0.296 ,1.0 ,0.769 ,0.237 ,0.243 ,0.304, 0.718 ,0.730],
[0.189 ,0.219 ,0.769 ,1.0 ,0.212 ,0.226 ,0.291, 0.681 ,0.661],
[0.461 ,0.479 ,0.237 ,0.212 ,1.0 ,0.520 ,0.514, 0.313 ,0.245],
[0.506 ,0.530 ,0.243 ,0.226 ,0.520 ,1.0 ,0.473, 0.348 ,0.290],
[0.408 ,0.425 ,0.304 ,0.291 ,0.514 ,0.473 ,1.0 , 0.374 ,0.306],
[0.280 ,0.311 ,0.718 ,0.681 ,0.313 ,0.348 ,0.374, 1.0 ,0.672],
[0.241 ,0.311 ,0.730 ,0.661 ,0.245 ,0.290 ,0.306, 0.672 ,1.0 ]])
self.lambdas = np.array([[0.70, 0.74, 0.39, 0.37, 0.65, 0.72, 0.60, 0.51, 0.48],
[-0.12,-0.08,0.81, 0.75,-0.03,-0.05, 0.09, 0.65, 0.67],
[0.15, 0.22, 0.33, 0.08, 0.00, 0.00, 0.00, 0.00, 0.00],
[0.00, 0.00, 0.00, 0.00, 0.37, 0.15, 0.35, 0.02,-0.12]])
self.taus = np.diag([0.48, 0.41, 0.09, 0.31, 0.44, 0.46, 0.52, 0.32, 0.32])
self.params_ex = {'lambdas' : self.lambdas, 'taus' : self.taus}
self.emmlfa_ex = EMMlFactorAnalysis(4,100,self.params_ex)
pattern = np.array([[1,1,1,1,1,1,1,1,1],
[1,1,1,1,1,1,1,1,1],
[1,1,1,1,0,0,0,0,0],
[0,0,0,0,1,1,1,1,1]])
self.params_cm = { 'lambdas' : self.lambdas, 'taus' : self.taus, 'factor_pattern' : pattern }
self.emmlfa_cm = EMMlFactorAnalysis(4,100,self.params_cm, fa_type="confirmatory")
def tearDown(self):
print "TEARDOWN"
def test_exploratory(self):
self.emmlfa_ex.fit(self.cyy)
l = self.emmlfa_ex.lambdas
t = self.emmlfa_ex.taus
assert(abs(np.sum(sig_tilt(l,t)) - np.sum(self.cyy)) < 1)
def test_woodbury(self):
res = woodbury(self.lambdas, self.taus)
true = linalg.inv(self.taus + np.transpose(self.lambdas).dot(self.lambdas))
assert_almost_equal(np.sum(res), np.sum(true), 10)
def test_confirmatory(self):
self.emmlfa_cm.fit(self.cyy)
l = self.emmlfa_cm.lambdas
t = self.emmlfa_cm.taus
assert(abs(np.sum(sig_tilt(l,t)) - np.sum(self.cyy)) < 1)
def test_varimax(self):
test = np.array([[-0.21855 , 0.908532, -0.32892 , 0.136455],
[ 0.841874, -0.36596 , -0.2355 , -0.31916 ],
[ 0.830432, 0.383946, -0.2225 , -0.33684 ],
[ 0.381244, 0.169304, -0.20553 , 0.885294],
[ 0.768423, 0.058657, 0.601702, 0.209855],
[ 0.361017, 0.036609, 0.928671, 0.076788],
[ 0.919046, 0.269187, -0.21876 , -0.18718 ],
[ 0.971137, 0.049841, 0.163809, 0.166061],
[ 0.334276, -0.82175 , -0.38167 , 0.25945 ]])
expect = np.array([[ 0.09129682, 0.88933586, -0.34325038, 0.28797331],
[ 0.80891817, -0.58100963, 0.05820452, -0.06850804],
[ 0.98631709, 0.14231445, 0.08083308, 0.01976065],
[ 0.10435783, -0.00513476, 0.0910481 , 0.99034953],
[ 0.40020067, -0.10117614, 0.87719403, 0.24521912],
[-0.00140417, 0.02477792, 0.99722012, -0.07025541],
[ 0.97714768, 0.00143377, 0.13723808, 0.1623131 ],
[ 0.72024453, -0.2037375 , 0.55082862, 0.36922558],
[ 0.10306332, -0.92631601, -0.20016431, 0.30207721]])
result = varimax(test)
assert(abs(np.sum(result) - np.sum(expect)) < 0.1)
