def test_block_entropy_base_4(self):
self.assertAlmostEqual(2.500000,
block_entropy([3, 3, 3, 2, 1, 0, 0, 0, 1], 2),
places=6)
self.assertAlmostEqual(2.405639,
block_entropy([2, 2, 3, 3, 3, 3, 2, 1, 0], 2),
places=6)
def test_block_entropy_base_2_ensemble(self):
xs = [[1, 1, 0, 0, 1, 0, 0, 1], [0, 0, 0, 1, 0, 0, 0, 1]]
self.assertAlmostEqual(1.788450, block_entropy(xs, 2), places=6)
xs = [[1, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 1, 1, 1, 1, 0, 0, 0],
[1, 0, 0, 0, 0, 0, 0, 1, 1], [1, 0, 0, 0, 0, 0, 0, 1, 1],
[0, 0, 0, 0, 0, 1, 1, 0, 0], [0, 0, 0, 0, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0, 0, 1, 1], [0, 0, 0, 1, 1, 1, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 1, 1, 0]]
self.assertAlmostEqual(1.649204, block_entropy(xs, 2), places=6)
def test_block_entropy_base_2(self):
self.assertAlmostEqual(1.950212,
block_entropy([1, 1, 0, 0, 1, 0, 0, 1],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(0.543564,
block_entropy([1, 0, 0, 0, 0, 0, 0, 0, 0],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.811278,
block_entropy([0, 0, 1, 1, 1, 1, 0, 0, 0],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.548795,
block_entropy([1, 0, 0, 0, 0, 0, 0, 1, 1],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.548795,
block_entropy([0, 0, 0, 0, 0, 1, 1, 0, 0],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.548795,
block_entropy([0, 0, 0, 0, 1, 1, 0, 0, 0],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.811278,
block_entropy([1, 1, 1, 0, 0, 0, 0, 1, 1],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.811278,
block_entropy([0, 0, 0, 1, 1, 1, 1, 0, 0],
2,
local=True).mean(),
places=6)
self.assertAlmostEqual(1.548795,
block_entropy([0, 0, 0, 0, 0, 0, 1, 1, 0],
2,
local=True).mean(),
places=6)
def changepoint_detection(ts, percent=0.05, plot=True, num_c=4):
length = len(ts)
bar = int(percent * length)
ts = np.array(ts)[bar:-bar]
min_val, model = length, "l1"
algo = rpt.Dynp(model="normal").fit(np.array(ts))
dic = {"best": [0, length]}
try:
for i in range(num_c):
my_bkps = algo.predict(n_bkps=i)
if plot:
rpt.show.display(np.array(ts), my_bkps, figsize=(10, 6))
plt.show()
start_borne, full_entro = 0, 0
for end_borne in my_bkps:
val = block_entropy(ts[start_borne:end_borne], k=1)
full_entro = val + full_entro
start_borne = end_borne
if full_entro == 0: break
elif full_entro < min_val:
min_val = full_entro
dic["best"] = [0] + my_bkps
else:
pass
except Exception as e:
print(e)
print("Not enough point")
return {"best": [0, length]}
return dic
def test_block_entropy_base_4_ensemble(self):
xs = [[3, 3, 3, 2, 1, 0, 0, 0, 1], [2, 2, 3, 3, 3, 3, 2, 1, 0],
[0, 0, 0, 0, 1, 1, 0, 0, 0], [1, 1, 0, 0, 0, 1, 1, 2, 2]]
self.assertAlmostEqual(3.010977, block_entropy(xs, 2), places=6)
def test_block_entropy_empty(self):
with self.assertRaises(ValueError):
block_entropy([], 1)
def test_block_entropy_negative_states(self):
with self.assertRaises(InformError):
block_entropy([-1, 0, 0], k=3)
def test_block_entropy_long_history(self):
with self.assertRaises(InformError):
block_entropy([1, 2], k=2)
with self.assertRaises(InformError):
block_entropy([1, 2], k=3)
def test_block_entropy_zero_history(self):
with self.assertRaises(InformError):
block_entropy([1, 2], k=0)
def test_block_entropy_short_series(self):
with self.assertRaises(InformError):
block_entropy([1], k=1)
def test_block_entropy_dimensions(self):
with self.assertRaises(ValueError):
block_entropy([[[1]]], 1)