def test_lnp_spikes_parametric_nonlinearity():
w_true, X, y, dims, dt = generate_spike_train(noise='white')
(X_train, y_train), (X_dev, y_dev), (_, _) = split_data(X,
y,
dt,
frac_train=0.8,
frac_dev=0.2)
df = [3, 4]
model = splineLNP(X_train,
y_train,
dims,
df=df,
dt=dt,
output_nonlinearity='spline')
model.initialize_parametric_nonlinearity(init_to='exponential',
params_dict={'df': 11})
model.fit(extra={
'X': X_dev,
'y': y_dev
},
num_iters=50,
metric='corrcoef',
beta=0.01,
verbose=False,
tolerance=10,
fit_nonlinearity=True)
assert model.output_nonlinearity == 'spline'
assert model.nl_basis is not None
assert model.nl_xrange is not None
assert model.nl_params is not None
assert model.fnl_fitted is not None
def test_splinelnp_spikes_split_data():
w_true, X, y, dims, dt = generate_spike_train(noise='white')
(X_train, y_train), (X_dev, y_dev), (_, _) = split_data(X,
y,
dt,
frac_train=0.8,
frac_dev=0.2)
df = [3, 4]
model = splineLNP(X_train,
y_train,
dims,
df,
dt=dt,
output_nonlinearity='exponential')
model.fit(extra={
'X': X_dev,
'y': y_dev
},
num_iters=50,
metric='corrcoef',
beta=0.01,
verbose=0,
tolerance=10)
assert model.w_spl.size == w_true.size
def test_lnp_spikes_history_filter():
w_true, X, y, dims, dt = generate_spike_train(noise='white')
(X_train, y_train), (X_dev, y_dev), (_, _) = split_data(X,
y,
dt,
frac_train=0.8,
frac_dev=0.2)
df = [3, 4]
model = splineLNP(X_train, y_train, dims, df=df, dt=dt)
model.initialize_history_filter(dims[0], df[0])
model.fit(extra={
'X': X_dev,
'y': y_dev
},
num_iters=50,
metric='corrcoef',
beta=0.01,
verbose=False,
tolerance=10,
fit_history_filter=True,
fit_nonlinearity=False)
assert model.h_opt is not None
assert model.h_spl is not None
def test_lnp_STC():
w_true, X, y, dims, dt = generate_spike_train(noise='white')
model = LNP(X, y, dims=dims, dt=dt)
model.fit_STC(n_repeats=100, verbose=10)
assert model.w_stc is not None
assert model.w_stc is not None
def test_lnp_spikes():
w_true, X, y, dims, dt = generate_spike_train(noise='white')
model = LNP(X, y, dims=dims, dt=dt)
model.fit(metric='corrcoef',
num_iters=100,
verbose=0,
tolerance=10,
beta=0.01)
assert model.w_opt.size == w_true.size
def test_splinelnp_spikes():
w_true, X, y, dims, dt = generate_spike_train(noise='white')
df = [3, 4]
model = splineLNP(X, y, dims=dims, dt=dt, df=df)
model.fit(metric='corrcoef',
num_iters=100,
verbose=0,
tolerance=10,
beta=0.01)
assert model.w_spl.size == w_true.size