Skip to main content
. 2026 Sep 9;26(18):5741. doi: 10.3390/s26185741
Algorithm A2. Sample entropy calculation
Input: Time series data X = [x1, x2, …, xN], Embedding dimension m, Tolerance r
Output: Sample entropy value SampEn
1. Define template matching function count_matches(X, m, r):
     N = length(X)
     count = 0
     FOR i = 1 TO N − m:
              FOR j = i + 1 TO N – m + 1:
                     Calculate Chebyshev distance:
                         dist = max (|X[i:i + m − 1] − X[j:j + m − 1]|)
                     IF dist < r:
                           count = count + 1
                     END IF
              END FOR
         END FOR
RETURN count
2. Calculate match counts:
     B = count_matches(X, m, r)
     A = count_matches(X, m + 1, r)
3. Calculate sample entropy:
     IF B > 0 AND A > 0:
             SampEn = − ln(A/B)
     ELSE:
             SampEn = NaN
     END IF
4. RETURN SampEn