k = 1while np.sum(s[0:k]) / s_sum < retained_variance: k += 1
s是一个numpy向量.谢谢！
import numpy as np ss = np.cumsum(s) # array with cumulative sumk = ss.searchsorted(retained_variance*s_sum) # exploit that ss is monotonically increasing
编辑：西蒙指出
k = np.cumsum(s).searchsorted(retained_variance*s_sum) + 1
是与问题对应的值.
