我在哪里可以找到一些 Numpy 方法的效率 O(n)？

Question

我正在做一个学校项目，他们询问了一些 Numpy 方法的效率 O(n)，我找不到它们。谁能告诉我在哪里可以找到这些？

示例方法如：

numpy.linspace(x,y,z) numpy.meshgrid(x,y) numpy.zeroes(x,y)

Answer 1

您可以简单地测量不同问题规模的执行时间，以估算时间复杂度，

• numpy.zeros(n)：non-deterministic
• numpy.meshgrid(x,y) : O(n**2)
• numpy.linspace(0, 1, n) : O(n**1.6)

例如，下面是测量numpy.meshgrid(x,y) 时间复杂度的代码，它也可以用于其他 numpy 函数，

In [1]: import numpy as np
   ...: from time import time
   ...: import matplotlib.pyplot as plt
   ...: from scipy.optimize import curve_fit
   ...: %matplotlib inline
   ...: 
   ...: def complexity_model(x, n, A, C):
   ...:     return A*x**n + C
   ...: 
   ...: problem_size = np.logspace(2, 4, 10)
   ...: 
   ...: res = []
   ...: for N in problem_size:
   ...:     x = np.linspace(0, 1, N)
   ...:     y = x.copy()
   ...:     
   ...:     t0 = time()
   ...:     np.meshgrid(x,y)
   ...:     dt = time() - t0
   ...:     res.append(dt)
   ...: 
   ...: nn = np.logspace(np.log10(problem_size.min()), np.log10(problem_size.max()), 100)  
   ...: 
   ...: time_to_solution = np.asarray(res)
   ...: fig, ax = plt.subplots(1,1)
   ...: ax.loglog(problem_size, time_to_solution, 'o-b')
   ...: 
   ...: mask = problem_size > 100 # ignore initial points
   ...: 
   ...: popt, _ = curve_fit(complexity_model, problem_size[mask],
   ...:                               time_to_solution[mask],
   ...:                               p0=(1.0, 1.0,  0.0) )
   ...: print(popt)
   ...: ax.loglog(nn, complexity_model(nn, *popt), '--k')
   ...: 
   ...: 
   ...: ax.set_xlabel('Problem size: N')
   ...: ax.set_ylabel('Time to solution 
  [  1.94816942e+00   1.40955397e-08  -7.33862899e-04]

给出以下曲线，

对于足够大的数组大小，numpy.meshgrid(x,y) 的时间复杂度为O(n**α)，α = 1.95 ≈ 2。