In [1]:

    

import pandas as pd
import numpy as np
import math
%matplotlib inline


    

In [2]:

    

# サイクルあたりのステップ数
steps_per_cycle = 50
# 生成するサイクル数
number_of_cycles = 100


    

In [3]:

    

df = pd.DataFrame(np.arange(steps_per_cycle * number_of_cycles + 1), columns=["t"])
df.head()


    

    
Out[3]:






  

    

      

      
t
    
  
  

    

      
0
      
0
    
    

      
1
      
1
    
    

      
2
      
2
    
    

      
3
      
3
    
    

      
4
      
4
    
  

In [4]:

    

df["sin_t"] = df.t.apply(lambda x: math.sin(x * (2 * math.pi / steps_per_cycle)))


    

In [5]:

    

df[["sin_t"]].plot()


    

    
Out[5]:





<matplotlib.axes._subplots.AxesSubplot at 0x111ba3208>






    

In [6]:

    

df[["sin_t"]].head(steps_per_cycle * 2).plot()


    

    
Out[6]:





<matplotlib.axes._subplots.AxesSubplot at 0x111bca710>






    

In [7]:

    

df["sin_t+1"] = df["sin_t"].shift(-1)


    

In [8]:

    

df.tail()


    

    
Out[8]:






  

    

      

      
t
      
sin_t
      
sin_t+1
    
  
  

    

      
4996
      
4996
      
-4.817537e-01
      
-3.681246e-01
    
    

      
4997
      
4997
      
-3.681246e-01
      
-2.486899e-01
    
    

      
4998
      
4998
      
-2.486899e-01
      
-1.253332e-01
    
    

      
4999
      
4999
      
-1.253332e-01
      
3.928773e-15
    
    

      
5000
      
5000
      
3.928773e-15
      
NaN
    
  

In [9]:

    

df.dropna(inplace=True)
df.tail()


    

    
Out[9]:






  

    

      

      
t
      
sin_t
      
sin_t+1
    
  
  

    

      
4995
      
4995
      
-0.587785
      
-4.817537e-01
    
    

      
4996
      
4996
      
-0.481754
      
-3.681246e-01
    
    

      
4997
      
4997
      
-0.368125
      
-2.486899e-01
    
    

      
4998
      
4998
      
-0.248690
      
-1.253332e-01
    
    

      
4999
      
4999
      
-0.125333
      
3.928773e-15
    
  

In [10]:

    

df[["sin_t", "sin_t+1"]].head(steps_per_cycle).plot()


    

    
Out[10]:





<matplotlib.axes._subplots.AxesSubplot at 0x111c642b0>






    

In [11]:

    

matrix = df[["sin_t", "sin_t+1"]].as_matrix()
matrix


    

    
Out[11]:





array([[  0.00000000e+00,   1.25333234e-01],
       [  1.25333234e-01,   2.48689887e-01],
       [  2.48689887e-01,   3.68124553e-01],
       ..., 
       [ -3.68124553e-01,  -2.48689887e-01],
       [ -2.48689887e-01,  -1.25333234e-01],
       [ -1.25333234e-01,   3.92877345e-15]])

In [12]:

    

np.save("normal.npy", matrix)


    

In [ ]: