Question

我尝试根据先前的问题进行矢量化处理，但仍然无法正常工作。

import numpy as np
import math


S0 = 50

k_list = np.linspace(S0 * 0.6, S0 * 1.4, 50)

K=k_list

d1 = np.vectorize(math.log(S0 / K))

print(d1)

Answer 1

In [141]: import math 
     ...:  
     ...:  
     ...: S0 = 50 
     ...:  
     ...: k_list = np.linspace(S0 * 0.6, S0 * 1.4, 50) 
     ...:  
     ...: K=k_list 
     ...:  
     ...: d1 = np.vectorize(math.log(S0 / K))                                          

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-141-5d2a0f276bc5> in <module>
      8 K=k_list
      9 
---> 10 d1 = np.vectorize(math.log(S0 / K))

TypeError: only size-1 arrays can be converted to Python scalars

np.vectorized参数不是函数，并且实际上会产生错误：

In [142]: math.log(S0 / K)                                                             
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-142-dedf1ab558ff> in <module>
----> 1 math.log(S0 / K)

TypeError: only size-1 arrays can be converted to Python scalars

log参数是一个数组。 math.log仅适用于1个数字，不适用于数组：

In [143]: S0 / K                                                                       
Out[143]: 
array([1.66666667, 1.62251656, 1.58064516, 1.5408805 , 1.50306748,
       1.46706587, 1.43274854, 1.4       , 1.36871508, 1.33879781,
       1.31016043, 1.28272251, 1.25641026, 1.23115578, 1.20689655,
       1.18357488, 1.16113744, 1.13953488, 1.11872146, 1.09865471,
       1.07929515, 1.06060606, 1.04255319, 1.0251046 , 1.00823045,
       0.99190283, 0.97609562, 0.96078431, 0.94594595, 0.93155894,
       0.917603  , 0.90405904, 0.89090909, 0.8781362 , 0.86572438,
       0.85365854, 0.8419244 , 0.83050847, 0.81939799, 0.80858086,
       0.7980456 , 0.78778135, 0.77777778, 0.76802508, 0.75851393,
       0.74923547, 0.74018127, 0.73134328, 0.72271386, 0.71428571])

np.log确实适用于数组输入：

In [145]: np.log(S0 / K)                                                               
Out[145]: 
array([ 0.51082562,  0.48397837,  0.45783309,  0.43235401,  0.40750801,
        0.3832644 ,  0.35959465,  0.33647224,  0.3138724 ,  0.29177206,
        0.27014959,  0.24898478,  0.22825865,  0.20795339,  0.18805223,
        0.16853942,  0.14940008,  0.13062018,  0.11218648,  0.09408644,
        0.07630819,  0.0588405 ,  0.0416727 ,  0.02479466,  0.00819677,
       -0.00813013, -0.02419473, -0.04000533, -0.05556985, -0.07089582,
       -0.08599045, -0.10086061, -0.11551289, -0.12995357, -0.14418869,
       -0.15822401, -0.17206506, -0.18571715, -0.19918536, -0.21247459,
       -0.22558954, -0.2385347 , -0.25131443, -0.26393289, -0.27639411,
       -0.28870196, -0.30086016, -0.31287232, -0.3247419 , -0.33647224])

使用vectorize的正确方法（如果有这样的事情:)）是：

In [146]: d1 = np.vectorize(lambda k: math.log(S0 / k))                                
In [147]: d1(K)                                                                        
Out[147]: 
array([ 0.51082562,  0.48397837,  0.45783309,  0.43235401,  0.40750801,
        0.3832644 ,  0.35959465,  0.33647224,  0.3138724 ,  0.29177206,
        0.27014959,  0.24898478,  0.22825865,  0.20795339,  0.18805223,
        0.16853942,  0.14940008,  0.13062018,  0.11218648,  0.09408644,
        0.07630819,  0.0588405 ,  0.0416727 ,  0.02479466,  0.00819677,
       -0.00813013, -0.02419473, -0.04000533, -0.05556985, -0.07089582,
       -0.08599045, -0.10086061, -0.11551289, -0.12995357, -0.14418869,
       -0.15822401, -0.17206506, -0.18571715, -0.19918536, -0.21247459,
       -0.22558954, -0.2385347 , -0.25131443, -0.26393289, -0.27639411,
       -0.28870196, -0.30086016, -0.31287232, -0.3247419 , -0.33647224])

np.vectorize用于TypeError：只能将size-1数组转换为Python标量

1 个答案: