Dynamic Time Warping 動態時間規整演算法python實現__演算法

找的參考文章給的代碼有點問題，調整了一下。具體原理就不贅述了，可參考這篇，實際使用也可以使用python的fastdtw包。

import sysdef distance(x,y):    return abs(x-y)def dtw():    X = [1, 2, 3, 4]    Y = [1, 2, 7, 4, 5]    M = [[distance(X[i], Y[i]) for i in range(len(X))] \    for j in range(len(Y))]    # print(M)    l1 = len(X)    l2 = len(Y)    D = [[0 for i in range(l1 + 1)] for i in range(l2 + 1)]    # D[0][0] = 0    for i in range(1, l1 + 1):        D[0][i] = sys.maxsize    for j in range(1, l2 + 1):        D[j][0] = sys.maxsize    for j in range(1, l2 + 1):        for i in range(1, l1 + 1):            D[j][i] = distance(X[i -1],Y[j-1 ])+ \                min(D[j - 1][i], D[j][i - 1], D[j - 1][i - 1])            print(j,i,D[j][i])    print(D)if __name__ == '__main__':    dtw()