def test_kfold_xval_MD_mod():
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv,
bvals_pv,
bvecs_pv,
mask_pv,
"decaying_exp",
10,
initial=-0.5,
bounds=[(None, None)],
signal="log")
# Check to see if the sum of the squared errors is around the values
# that we want.
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv,
bvals_pv,
bvecs_pv,
mask_pv,
"decaying_exp_plus_const",
10,
initial=(-0.5, -0.5),
bounds=[(None, None),
(None, None)],
signal="log")
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv,
bvals_pv,
bvecs_pv,
mask_pv,
"two_decaying_exp",
10,
initial=(-0.5, -0.5, -0.5),
bounds=[(None, None),
(None, None),
(None, None)],
signal="log")
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "single_exp_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "single_exp_nf_rs",
10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "bi_exp_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "bi_exp_nf_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
def test_kfold_xval_MD_mod():
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "decaying_exp",
10, initial = -0.5,
bounds=[(None, None)],
signal="log")
# Check to see if the sum of the squared errors is around the values
# that we want.
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "decaying_exp_plus_const",
10, initial = (-0.5, -0.5),
bounds=[(None, None), (None, None)],
signal="log")
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "two_decaying_exp",
10, initial=(-0.5, -0.5, -0.5),
bounds=[(None, None), (None, None),
(None, None)],
signal="log")
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "single_exp_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "single_exp_nf_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "bi_exp_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
ss_err, predict_out = mdm.kfold_xval_MD_mod(data_pv, bvals_pv, bvecs_pv,
mask_pv, "bi_exp_nf_rs", 10)
npt.assert_equal(np.mean(ss_err) < 200, 1)
bvals = np.loadtxt(os.path.join(data_path, "bvals"))
bvecs = np.loadtxt(os.path.join(data_path, "bvecs"))
low = i*2000
# Make sure not to go over the edge of the mask:
high = np.min([(i+1) * 2000, int(np.sum(mask_data))])
# Now set the mask:
mask = np.zeros(wm_data_file.shape)
mask[mask_idx[0][low:high], mask_idx[1][low:high], mask_idx[2][low:high]] = 1
if im == "bi_exp_rs":
shorthand_im = "be"
elif im == "single_exp_rs":
shorthand_im = "se"
param_out, fit_out, _ = mdm.optimize_MD_params(data, bvals, bvecs, mask,
im, signal = "relative_signal")
cod, predict_out = mdm.kfold_xval_MD_mod(data, bvals, bvecs, mask,
im, 10, signal="relative_signal")
np.save(os.path.join(data_path, "im_cod_%s%s.npy"%(shorthand_im, i)), cod)
np.save(os.path.join(data_path, "im_predict_out_%s%s.npy"%(shorthand_im, i)), predict_out)
np.save(os.path.join(data_path, "im_param_out_%s%s.npy"%(shorthand_im, i)), param_out)
t2 = time.time()
print "This program took %4.2f minutes to run."%((t2 - t1)/60.)
