Last active
April 27, 2022 14:37
-
-
Save nnao45/ffc0909c6b2e986b14dfbd940dd9e2d8 to your computer and use it in GitHub Desktop.
Distribution Entropy (Peng Li, Chengyu Liu, Ke Li, Dingchang Zheng, Changchun Liu,Yinglong Hou. "Assessing the complexity of short-term heartbeat interval series by distribution entropy", Med Biol Eng Comput,2015, 53: 77-87.)
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
""" | |
Ref. | |
[1] Peng Li, Chengyu Liu, Ke Li, Dingchang Zheng, Changchun Liu, | |
Yinglong Hou. "Assessing the complexity of short-term heartbeat | |
interval series by distribution entropy", Med Biol Eng Comput, | |
2015, 53: 77-87. | |
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ | |
(C) Peng Li 2013-2017 | |
If you use the code, please make sure that you cite reference [1] | |
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ | |
""" | |
import numpy as np | |
import pandas as pd | |
from scipy.linalg import hankel | |
from scipy.spatial.distance import pdist | |
from typing import List | |
import sys | |
import math | |
def disten(ser: List[float], m: int = 2, tau: int = 8 , B: int = 512) -> float: | |
""" | |
@param ser: time-series (vector in a column) | |
@param m: embedding dimension (scalar) | |
@param tau: time delay (scalar) | |
@param B: bin number for histogram (scalar) | |
""" | |
# rescaling | |
rescaled = [y / (max(ser) - min(ser) + sys.float_info.epsilon) for y in [x - min(ser) for x in ser]] | |
# distance matrix | |
N = len(rescaled) - (m - 1) * tau | |
if N < 0: | |
raise(f"ser is too short: {len(ser)}") | |
ind = hankel(np.arange(1, N+1), np.arange(N, len(rescaled)+1)) | |
rnt = [[rescaled[z-1] for z in y] for y in [x[::tau] for x in ind]] | |
dv = pdist(rnt, 'chebychev') | |
# esimating probability density by histogram | |
num = pd.cut(dv, np.linspace(0, 1, B), include_lowest=True).value_counts().to_numpy() | |
freq = [x / num.sum() for x in num] | |
# disten calculation | |
prepared = [math.log2(y) for y in [x + sys.float_info.epsilon for x in freq]] | |
return -sum([x * y for (x, y) in zip(prepared, freq)]) / math.log2(B) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment