def test_score_1():
grid = NDGrid(n_bins_per_feature=5, min=-np.pi, max=np.pi)
trajs = DoubleWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajs)
model = (ContinuousTimeMSM(verbose=False, lag_time=10, n_timescales=3)
.fit(seqs))
np.testing.assert_approx_equal(model.score(seqs), model.score_)
def test_uncertainties_backward():
n = 4
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(verbose=False).fit(seqs)
sigma_ts = model.uncertainty_timescales()
sigma_lambda = model.uncertainty_eigenvalues()
sigma_pi = model.uncertainty_pi()
sigma_K = model.uncertainty_K()
yield lambda: np.testing.assert_array_almost_equal(
sigma_ts, [9.13698928, 0.12415533, 0.11713719])
yield lambda: np.testing.assert_array_almost_equal(
sigma_lambda, [1.76569687e-19, 7.14216858e-05, 3.31210649e-04, 3.55556718e-04])
yield lambda: np.testing.assert_array_almost_equal(
sigma_pi, [0.00741467, 0.00647945, 0.00626743, 0.00777847])
yield lambda: np.testing.assert_array_almost_equal(
sigma_K,
[[ 3.39252419e-04, 3.39246173e-04, 0.00000000e+00, 1.62090239e-06],
[ 3.52062861e-04, 3.73305510e-04, 1.24093936e-04, 0.00000000e+00],
[ 0.00000000e+00, 1.04708186e-04, 3.45098923e-04, 3.28820213e-04],
[ 1.25455972e-06, 0.00000000e+00, 2.90118599e-04, 2.90122944e-04]])
yield lambda: np.testing.assert_array_almost_equal(
model.ratemat_,
[[ -2.54439564e-02, 2.54431791e-02, 0.00000000e+00, 7.77248586e-07],
[ 2.64044208e-02,-2.97630373e-02, 3.35861646e-03, 0.00000000e+00],
[ 0.00000000e+00, 2.83988103e-03, -3.01998380e-02, 2.73599570e-02],
[ 6.01581838e-07, 0.00000000e+00, 2.41326592e-02, -2.41332608e-02]])
def test_score_3():
import warnings
warnings.simplefilter('ignore')
from msmbuilder.example_datasets.muller import MULLER_PARAMETERS as PARAMS
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
ds = MullerPotential(random_state=0).get()['trajectories']
assignments = cluster.fit_transform(ds)
train_indices = [9, 4, 3, 6, 2]
test_indices = [8, 0, 5, 7, 1]
model = ContinuousTimeMSM(lag_time=3,
n_timescales=1,
sliding_window=False,
ergodic_cutoff=1)
train_data = [assignments[i] for i in train_indices]
test_data = [assignments[i] for i in test_indices]
model.fit(train_data)
train = model.score_
test = model.score(test_data)
print(train, test)
def test_score_3():
ds = MullerPotential(random_state=0).get_cached().trajectories
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
train_indices = [9, 4, 3, 6, 2]
test_indices = [8, 0, 5, 7, 1]
temp = '1.0929e-02 5.4147e-02 9.8362e-02 0.1000e+00 6.0455e-02 2.8775e-02\
6.6456e-02 3.3957e-02 4.1484e-03 0.1000e+00 5.0847e-02 1.1516e-02\
3.5266e-02 1.2830e-02 0.1000e+00 2.1801e-02 1.6639e-02 9.4932e-03\
0.1000e+00 0.1000e+00 1.1050e-01 4.0076e-03 0.1000e+00 0.1000e+00\
1.8930e-02 -7.1060e+00 -4.5787e+00 -2.4950e+00 -4.0964e+00 -7.4127e+00\
-6.7574e+00 -4.7137e+00 -3.9530e+00 -4.5781e+00 -7.4585e+00 -6.4634e+00\
-5.8060e+00 -5.4783e+00 -5.3519e+00 -7.4653e+00 -6.5113e+00 -2.1477e+00\
-4.8138e+00 -9.7187e+00 -9.0358e+00 -1.4599e+00 -8.8985e-01 -8.3461e+00\
-7.0930e+00 -2.7618e+00 -6.7421e+00'
model = PESContinuousTimeMSM(lag_time=3, n_timescales=1, sliding_window=False,
ergodic_cutoff=1)
model.theta_ = list(map(np.float64, temp.split()))
train_data = [assignments[i] for i in train_indices]
test_data = [assignments[i] for i in test_indices]
model.fit(train_data)
print(model.summarize())
train = model.score_
test = model.score(test_data)
print(train, test)
def test_5():
grid = NDGrid(n_bins_per_feature=2)
seqs = grid.fit_transform(load_quadwell(random_state=0)['trajectories'])
model2 = BayesianContinuousTimeMSM(n_samples=100).fit(seqs)
print(model2.summarize())
def test_uncertainties_backward():
n = 4
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
trajs = DoubleWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajs)
model = PESContinuousTimeMSM(verbose=False).fit(seqs)
sigma_ts = model.uncertainty_timescales()
sigma_lambda = model.uncertainty_eigenvalues()
sigma_pi = model.uncertainty_pi()
sigma_K = model.uncertainty_K()
yield lambda: np.testing.assert_array_almost_equal(
sigma_ts, [9.508936, 0.124428, 0.117638])
yield lambda: np.testing.assert_array_almost_equal(
sigma_lambda,
[1.76569687e-19, 7.14216858e-05, 3.31210649e-04, 3.55556718e-04])
yield lambda: np.testing.assert_array_almost_equal(
sigma_pi, [0.007496, 0.006564, 0.006348, 0.007863])
yield lambda: np.testing.assert_array_almost_equal(
sigma_K,
[[0.000339, 0.000339, 0., 0.],
[0.000352, 0.000372, 0.000122, 0.],
[0., 0.000103, 0.000344, 0.000329],
[0., 0., 0.00029, 0.00029]])
yield lambda: np.testing.assert_array_almost_equal(
model.ratemat_,
[[-0.0254, 0.0254, 0., 0.],
[0.02636, -0.029629, 0.003269, 0.],
[0., 0.002764, -0.030085, 0.027321],
[0., 0., 0.024098, -0.024098]])
def test_uncertainties_backward():
n = 4
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(verbose=False).fit(seqs)
sigma_ts = model.uncertainty_timescales()
sigma_lambda = model.uncertainty_eigenvalues()
sigma_pi = model.uncertainty_pi()
sigma_K = model.uncertainty_K()
yield lambda: np.testing.assert_array_almost_equal(
sigma_ts, [9.13698928, 0.12415533, 0.11713719])
yield lambda: np.testing.assert_array_almost_equal(sigma_lambda, [
1.76569687e-19, 7.14216858e-05, 3.31210649e-04, 3.55556718e-04
])
yield lambda: np.testing.assert_array_almost_equal(
sigma_pi, [0.00741467, 0.00647945, 0.00626743, 0.00777847])
yield lambda: np.testing.assert_array_almost_equal(sigma_K, [
[3.39252419e-04, 3.39246173e-04, 0.00000000e+00, 1.62090239e-06],
[3.52062861e-04, 3.73305510e-04, 1.24093936e-04, 0.00000000e+00],
[0.00000000e+00, 1.04708186e-04, 3.45098923e-04, 3.28820213e-04],
[1.25455972e-06, 0.00000000e+00, 2.90118599e-04, 2.90122944e-04]
])
yield lambda: np.testing.assert_array_almost_equal(model.ratemat_, [
[-2.54439564e-02, 2.54431791e-02, 0.00000000e+00, 7.77248586e-07],
[2.64044208e-02, -2.97630373e-02, 3.35861646e-03, 0.00000000e+00],
[0.00000000e+00, 2.83988103e-03, -3.01998380e-02, 2.73599570e-02],
[6.01581838e-07, 0.00000000e+00, 2.41326592e-02, -2.41332608e-02]
])
def test_5():
grid = NDGrid(n_bins_per_feature=2)
seqs = grid.fit_transform(load_quadwell(random_state=0)['trajectories'])
model2 = BayesianContinuousTimeMSM(n_samples=100).fit(seqs)
print(model2.summarize())
def test_score_2():
ds = MullerPotential(random_state=0).get_cached().trajectories
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
test_indices = [5, 0, 4, 1, 2]
train_indices = [3, 6, 7, 8, 9]
model = PESContinuousTimeMSM(lag_time=3, n_timescales=1)
temp='3.13268076e-02 1.45678356e-01 3.12558665e-01 0.10000000e+00\
2.72951862e-02 9.64773504e-02 1.26398091e-01 1.27775726e-01\
2.74208403e-03 9.57265955e-04 1.89498433e-01 0.10010000e+00\
9.73644477e-02 4.16877008e-02 0.10010000e+00 1.73803374e-01\
4.31281724e-02 0.10001000e+00 0.10010000e+00 4.03763450e-01\
0.10001000e+00 0.10001000e+00 0.10010000e+00 2.78156537e-01\
-7.75852152e+00 -4.77716045e+00 -2.67428479e+00 -4.33901900e+00\
-9.23925293e+00 -6.65216281e+00 -4.88309143e+00 -4.04247463e+00\
-4.67140081e+00 -7.95471679e+00 -6.26342874e+00 -6.02515423e+00\
-5.64532492e+00 -5.56770596e+00 -7.66164067e+00 -6.22050765e+00\
-2.12577068e+00 -4.84152585e+00 -9.21360166e+00 -1.43207874e+00\
-8.55459835e-01 -9.21329384e+00 -6.99418825e+00 -2.73060233e+00\
-6.60364249e+00'
model.theta_= list(map(np.float64, temp.split()))
model.fit([assignments[i] for i in train_indices])
print('Initial theta: \n')
print(model._initial_guess(model.countsmat_))
test = model.score([assignments[i] for i in test_indices])
train = model.score_
print('train', train, 'test', test)
print(model.optimizer_state_)
# print(model.summarize())
assert 1 <= test < 2
assert 1 <= train < 2
def test_score_1():
grid = NDGrid(n_bins_per_feature=5, min=-np.pi, max=np.pi)
trajs = DoubleWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajs)
model = (PESContinuousTimeMSM(verbose=False, lag_time=10, n_timescales=3)
.fit(seqs))
np.testing.assert_approx_equal(model.score(seqs), model.score_)
def test_uncertainties_backward():
n = 4
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
trajs = DoubleWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajs)
model = ContinuousTimeMSM(verbose=False).fit(seqs)
sigma_ts = model.uncertainty_timescales()
sigma_lambda = model.uncertainty_eigenvalues()
sigma_pi = model.uncertainty_pi()
sigma_K = model.uncertainty_K()
yield lambda: np.testing.assert_array_almost_equal(
sigma_ts, [9.508936, 0.124428, 0.117638])
yield lambda: np.testing.assert_array_almost_equal(
sigma_lambda,
[1.76569687e-19, 7.14216858e-05, 3.31210649e-04, 3.55556718e-04])
yield lambda: np.testing.assert_array_almost_equal(
sigma_pi, [0.007496, 0.006564, 0.006348, 0.007863])
yield lambda: np.testing.assert_array_almost_equal(
sigma_K,
[[0.000339, 0.000339, 0., 0.],
[0.000352, 0.000372, 0.000122, 0.],
[0., 0.000103, 0.000344, 0.000329],
[0., 0., 0.00029, 0.00029]])
yield lambda: np.testing.assert_array_almost_equal(
model.ratemat_,
[[-0.0254, 0.0254, 0., 0.],
[0.02636, -0.029629, 0.003269, 0.],
[0., 0.002764, -0.030085, 0.027321],
[0., 0., 0.024098, -0.024098]])
def test_optimize_1():
n = 100
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(use_sparse=True, verbose=True).fit(seqs)
y, x, n = model.loglikelihoods_.T
x = x-x[0]
cross = np.min(np.where(n==n[-1])[0])
def test_doublewell():
trjs = load_doublewell(random_state=0)['trajectories']
for n_states in [10, 50]:
clusterer = NDGrid(n_bins_per_feature=n_states)
assignments = clusterer.fit_transform(trjs)
for sliding_window in [True, False]:
model = ContinuousTimeMSM(lag_time=100, sliding_window=sliding_window)
model.fit(assignments)
assert model.optimizer_state_.success
def test_doublewell():
trjs = load_doublewell(random_state=0)['trajectories']
for n_states in [10, 50]:
clusterer = NDGrid(n_bins_per_feature=n_states)
assignments = clusterer.fit_transform(trjs)
for sliding_window in [True, False]:
model = ContinuousTimeMSM(lag_time=100,
sliding_window=sliding_window)
model.fit(assignments)
assert model.optimizer_state_.success
def test_doublewell():
trjs = DoubleWell(random_state=0).get_cached().trajectories
for n_states in [10, 36]:
clusterer = NDGrid(n_bins_per_feature=n_states)
assignments = clusterer.fit_transform(trjs)
for sliding_window in [True, False]:
model = PESContinuousTimeMSM(lag_time=100,
sliding_window=sliding_window)
model.fit(assignments)
# print(model.summarize())
print(model.optimizer_state_)
assert model.optimizer_state_.success
def test_hessian():
grid = NDGrid(n_bins_per_feature=10, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
seqs = [seqs[i] for i in range(10)]
lag_time = 10
model = ContinuousTimeMSM(verbose=True, lag_time=lag_time)
model.fit(seqs)
msm = MarkovStateModel(verbose=False, lag_time=lag_time)
print(model.summarize())
print('MSM timescales\n', msm.fit(seqs).timescales_)
print('Uncertainty K\n', model.uncertainty_K())
print('Uncertainty pi\n', model.uncertainty_pi())
def test_hessian_3():
grid = NDGrid(n_bins_per_feature=4, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
seqs = [seqs[i] for i in range(10)]
lag_time = 10
model = ContinuousTimeMSM(verbose=False, lag_time=lag_time)
model.fit(seqs)
msm = MarkovStateModel(verbose=False, lag_time=lag_time)
print(model.summarize())
# print('MSM timescales\n', msm.fit(seqs).timescales_)
print('Uncertainty K\n', model.uncertainty_K())
print('Uncertainty eigs\n', model.uncertainty_eigenvalues())
def test_hessian_3():
grid = NDGrid(n_bins_per_feature=4, min=-np.pi, max=np.pi)
trajs = DoubleWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajs)
seqs = [seqs[i] for i in range(10)]
lag_time = 10
model = ContinuousTimeMSM(verbose=False, lag_time=lag_time)
model.fit(seqs)
msm = MarkovStateModel(verbose=False, lag_time=lag_time)
print(model.summarize())
# print('MSM timescales\n', msm.fit(seqs).timescales_)
print('Uncertainty K\n', model.uncertainty_K())
print('Uncertainty eigs\n', model.uncertainty_eigenvalues())
def test_5():
trjs = DoubleWell(random_state=0).get_cached().trajectories
clusterer = NDGrid(n_bins_per_feature=5)
mle_msm = MarkovStateModel(lag_time=100, verbose=False)
b_msm = BayesianMarkovStateModel(lag_time=100, n_samples=1000, n_chains=8, n_steps=1000, random_state=0)
states = clusterer.fit_transform(trjs)
b_msm.fit(states)
mle_msm.fit(states)
# this is a pretty silly test. it checks that the mean transition
# matrix is not so dissimilar from the MLE transition matrix.
# This shouldn't necessarily be the case anyways -- the likelihood is
# not "symmetric". And the cutoff chosen is just heuristic.
assert np.linalg.norm(b_msm.all_transmats_.mean(axis=0) - mle_msm.transmat_) < 1e-2
def test_hessian_1():
n = 5
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(use_sparse=False).fit(seqs)
theta = model.theta_
C = model.countsmat_
hessian1 = _ratematrix.hessian(theta, C, n)
Hfun = nd.Jacobian(lambda x: _ratematrix.loglikelihood(x, C, n)[1])
hessian2 = Hfun(theta)
# not sure what the cutoff here should be (see plot_test_hessian)
assert np.linalg.norm(hessian1-hessian2) < 1
def test_score_2():
ds = MullerPotential(random_state=0).get_cached().trajectories
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
test_indices = [5, 0, 4, 1, 2]
train_indices = [3, 6, 7, 8, 9]
model = ContinuousTimeMSM(lag_time=3, n_timescales=1)
model.fit([assignments[i] for i in train_indices])
test = model.score([assignments[i] for i in test_indices])
train = model.score_
print('train', train, 'test', test)
assert 1 <= test < 2
assert 1 <= train < 2
def test_guess():
ds = MullerPotential(random_state=0).get_cached().trajectories
cluster = NDGrid(n_bins_per_feature=5,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
model1 = ContinuousTimeMSM(guess='log')
model1.fit(assignments)
model2 = ContinuousTimeMSM(guess='pseudo')
model2.fit(assignments)
diff = model1.loglikelihoods_[-1] - model2.loglikelihoods_[-1]
assert np.abs(diff) < 1e-3
assert np.max(np.abs(model1.ratemat_ - model2.ratemat_)) < 1e-1
def test_fit_2():
grid = NDGrid(n_bins_per_feature=5, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(verbose=False, lag_time=10)
model.fit(seqs)
t1 = np.sort(model.timescales_)
t2 = -1 / np.sort(np.log(np.linalg.eigvals(model.transmat_))[1:])
model = MarkovStateModel(verbose=False, lag_time=10)
model.fit(seqs)
t3 = np.sort(model.timescales_)
np.testing.assert_array_almost_equal(t1, t2)
# timescales should be similar to MSM (withing 50%)
assert abs(t1[-1] - t3[-1]) / t1[-1] < 0.50
def test_fit_2():
grid = NDGrid(n_bins_per_feature=5, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(verbose=True, lag_time=10)
model.fit(seqs)
t1 = np.sort(model.timescales_)
t2 = -1/np.sort(np.log(np.linalg.eigvals(model.transmat_))[1:])
model = MarkovStateModel(verbose=False, lag_time=10)
model.fit(seqs)
t3 = np.sort(model.timescales_)
np.testing.assert_array_almost_equal(t1, t2)
# timescales should be similar to MSM (withing 50%)
assert abs(t1[-1] - t3[-1]) / t1[-1] < 0.50
def test_score_2():
from msmbuilder.example_datasets.muller import MULLER_PARAMETERS as PARAMS
ds = MullerPotential(random_state=0).get()['trajectories']
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
test_indices = [5, 0, 4, 1, 2]
train_indices = [3, 6, 7, 8, 9]
model = ContinuousTimeMSM(lag_time=3, n_timescales=1)
model.fit([assignments[i] for i in train_indices])
test = model.score([assignments[i] for i in test_indices])
train = model.score_
print('train', train, 'test', test)
assert 1 <= test < 2
assert 1 <= train < 2
def test_score_2():
from msmbuilder.example_datasets.muller import MULLER_PARAMETERS as PARAMS
ds = MullerPotential(random_state=0).get()['trajectories']
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
test_indices = [5, 0, 4, 1, 2]
train_indices = [3, 6, 7, 8, 9]
model = ContinuousTimeMSM(lag_time=3, n_timescales=1)
model.fit([assignments[i] for i in train_indices])
test = model.score([assignments[i] for i in test_indices])
train = model.score_
print('train', train, 'test', test)
assert 1 <= test < 2
assert 1 <= train < 2
def test_guess():
from msmbuilder.example_datasets.muller import MULLER_PARAMETERS as PARAMS
cluster = NDGrid(n_bins_per_feature=5,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
ds = MullerPotential(random_state=0).get()['trajectories']
assignments = cluster.fit_transform(ds)
model1 = ContinuousTimeMSM(guess='log')
model1.fit(assignments)
model2 = ContinuousTimeMSM(guess='pseudo')
model2.fit(assignments)
assert np.abs(model1.loglikelihoods_[-1] - model2.loglikelihoods_[-1]) < 1e-3
assert np.max(np.abs(model1.ratemat_ - model2.ratemat_)) < 1e-1
def test_guess():
from msmbuilder.example_datasets.muller import MULLER_PARAMETERS as PARAMS
cluster = NDGrid(n_bins_per_feature=5,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
ds = MullerPotential(random_state=0).get()['trajectories']
assignments = cluster.fit_transform(ds)
model1 = ContinuousTimeMSM(guess='log')
model1.fit(assignments)
model2 = ContinuousTimeMSM(guess='pseudo')
model2.fit(assignments)
diff = model1.loglikelihoods_[-1] - model2.loglikelihoods_[-1]
assert np.abs(diff) < 1e-3
assert np.max(np.abs(model1.ratemat_ - model2.ratemat_)) < 1e-1
def test_5():
trjs = DoubleWell(random_state=0).get_cached().trajectories
clusterer = NDGrid(n_bins_per_feature=5)
mle_msm = MarkovStateModel(lag_time=100, verbose=False)
b_msm = BayesianMarkovStateModel(lag_time=100,
n_samples=1000,
n_chains=8,
n_steps=1000,
random_state=0)
states = clusterer.fit_transform(trjs)
b_msm.fit(states)
mle_msm.fit(states)
# this is a pretty silly test. it checks that the mean transition
# matrix is not so dissimilar from the MLE transition matrix.
# This shouldn't necessarily be the case anyways -- the likelihood is
# not "symmetric". And the cutoff chosen is just heuristic.
assert np.linalg.norm(
b_msm.all_transmats_.mean(axis=0) - mle_msm.transmat_) < 1e-2
def test_score_3():
ds = MullerPotential(random_state=0).get_cached().trajectories
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
assignments = cluster.fit_transform(ds)
train_indices = [9, 4, 3, 6, 2]
test_indices = [8, 0, 5, 7, 1]
model = ContinuousTimeMSM(lag_time=3, n_timescales=1, sliding_window=False,
ergodic_cutoff=1)
train_data = [assignments[i] for i in train_indices]
test_data = [assignments[i] for i in test_indices]
model.fit(train_data)
train = model.score_
test = model.score(test_data)
print(train, test)
def _plot_test_hessian():
# plot the difference between the numerical hessian and the analytic
# approximate hessian (opens Matplotlib window)
n = 5
grid = NDGrid(n_bins_per_feature=n, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(use_sparse=False).fit(seqs)
theta = model.theta_
C = model.countsmat_
hessian1 = _ratematrix.hessian(theta, C, n)
Hfun = nd.Jacobian(lambda x: _ratematrix.loglikelihood(x, C, n)[1])
hessian2 = Hfun(theta)
import matplotlib.pyplot as pp
pp.scatter(hessian1.flat, hessian2.flat, marker='x')
pp.plot(pp.xlim(), pp.xlim(), 'k')
print('Plotting...', file=sys.stderr)
pp.show()
def test_score_3():
from msmbuilder.example_datasets.muller import MULLER_PARAMETERS as PARAMS
cluster = NDGrid(n_bins_per_feature=6,
min=[PARAMS['MIN_X'], PARAMS['MIN_Y']],
max=[PARAMS['MAX_X'], PARAMS['MAX_Y']])
ds = MullerPotential(random_state=0).get()['trajectories']
assignments = cluster.fit_transform(ds)
train_indices = [9, 4, 3, 6, 2]
test_indices = [8, 0, 5, 7, 1]
model = ContinuousTimeMSM(lag_time=3, n_timescales=1, sliding_window=False, ergodic_cutoff=1)
train_data = [assignments[i] for i in train_indices]
test_data = [assignments[i] for i in test_indices]
model.fit(train_data)
train = model.score_
test = model.score(test_data)
print(train, test)
def get_dtrajs(X, xmin, xmax, m):
cluster = NDGrid(min=xmin, max=xmax, n_bins_per_feature=m)
dtrajs = cluster.fit_transform(X)
return dtrajs
def test_5():
grid = NDGrid(n_bins_per_feature=2)
trajectories = QuadWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajectories)
model2 = BayesianContinuousTimeMSM(n_samples=100).fit(seqs)
def test_5():
grid = NDGrid(n_bins_per_feature=2)
trajectories = QuadWell(random_state=0).get_cached().trajectories
seqs = grid.fit_transform(trajectories)
model2 = BayesianContinuousTimeMSM(n_samples=100).fit(seqs)
def test_score_1():
grid = NDGrid(n_bins_per_feature=5, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = ContinuousTimeMSM(verbose=False, lag_time=10, n_timescales=3).fit(seqs)
np.testing.assert_approx_equal(model.score(seqs), model.score_)
def test_score_1():
grid = NDGrid(n_bins_per_feature=5, min=-np.pi, max=np.pi)
seqs = grid.fit_transform(load_doublewell(random_state=0)['trajectories'])
model = (ContinuousTimeMSM(verbose=False, lag_time=10,
n_timescales=3).fit(seqs))
np.testing.assert_approx_equal(model.score(seqs), model.score_)
