def test_normality():
# Generate some 4-d random uniform data.
# The first 3 dimensions are like voxels, the last like time.
np.random.seed(159)
sim_resids = np.random.rand(2, 2, 2, 200)
# Force one of the time courses to be standard normal.
sim_resids[0,0,0] = np.random.randn(200)
# Do Shaprio-Wilk and Kruskal-Wallis
sw_3d = check_sw(sim_resids)
kw_3d = check_kw(sim_resids)
assert(sw_3d[0,0,0] > 0.05)
assert(sw_3d[1,0,0] < 0.05)
assert(kw_3d[0,0,0] > 0.05)
def test_normality():
# Generate some 4-d random uniform data.
# The first 3 dimensions are like voxels, the last like time.
np.random.seed(159)
sim_resids = np.random.rand(2, 2, 2, 200)
# Force one of the time courses to be standard normal.
sim_resids[0, 0, 0] = np.random.randn(200)
# Do Shaprio-Wilk and Kruskal-Wallis
sw_3d = check_sw(sim_resids)
kw_3d = check_kw(sim_resids)
assert (sw_3d[0, 0, 0] > 0.05)
assert (sw_3d[1, 0, 0] < 0.05)
assert (kw_3d[0, 0, 0] > 0.05)
def test_normality():
# Generate some 4-d random uniform data.
# The first 3 dimensions are like voxels, the last like time.
np.random.seed(159)
sim_resids = np.random.rand(2, 2, 2, 200)
# Force one of the time courses to be standard normal.
sim_resids[0,0,0] = np.random.randn(200)
# Do Shaprio-Wilk.
sw_3d = check_sw(sim_resids) # 4-d residuals, 3-d p-values
sw_1d = check_sw_masked(sim_resids.reshape((-1, sim_resids.shape[-1]))) # 2-d residuals, 1-d p-values
# Do Kruskal-Wallis.
kw_3d = check_kw(sim_resids)
assert(sw_3d[0,0,0] > 0.05)
assert(sw_3d[1,0,0] < 0.05)
# Two Shaprio-Wilk functions should do the same thing over arrays of different dimensions.
assert(sw_3d[0,0,0] == sw_1d[0])
assert(kw_3d[0,0,0] > 0.05)
def test_normality():
# Generate some 4-d random uniform data.
# The first 3 dimensions are like voxels, the last like time.
np.random.seed(159)
sim_resids = np.random.rand(2, 2, 2, 200)
# Force one of the time courses to be standard normal.
sim_resids[0, 0, 0] = np.random.randn(200)
# Do Shaprio-Wilk.
sw_3d = check_sw(sim_resids) # 4-d residuals, 3-d p-values
sw_1d = check_sw_masked(sim_resids.reshape(
(-1, sim_resids.shape[-1]))) # 2-d residuals, 1-d p-values
# Do Kruskal-Wallis.
kw_3d = check_kw(sim_resids)
assert (sw_3d[0, 0, 0] > 0.05)
assert (sw_3d[1, 0, 0] < 0.05)
# Two Shaprio-Wilk functions should do the same thing over arrays of different dimensions.
assert (sw_3d[0, 0, 0] == sw_1d[0])
assert (kw_3d[0, 0, 0] > 0.05)