In [1]:

    

import numpy as np
import pandas as pd

import statsmodels.api as sm
import statsmodels.formula.api as smf

import matplotlib.pyplot as plt
# import seaborn as sns
%matplotlib inline


    

In [2]:

    

data = pd.read_csv('../benchMarkingResult.txt',
                   header=None,
                   sep='\t',
                   names=('iteration',
                          'basic_result',
                          'efficient_result'))


    

In [3]:

    

x = data.iteration
y1 = data.basic_result
y2 = data.efficient_result


    

In [4]:

    

fig, ax = plt.subplots(figsize=(10,10))
ax.scatter(x, y1, s=15)
ax.scatter(x, y2, s=15, c='g')


    

    
Out[4]:





<matplotlib.collections.PathCollection at 0x11a1beb50>






    

In [5]:

    

d1 = {'x':x, 'y_basic':y1}
d2 = {'x':x, 'y_efficient':y2}


    

In [6]:

    

mod1 = smf.ols(formula='y_basic ~ x - 1',
               data = d1).fit()


    

In [7]:

    

print mod1.summary()


    

    




                            OLS Regression Results                            
==============================================================================
Dep. Variable:                y_basic   R-squared:                       0.993
Model:                            OLS   Adj. R-squared:                  0.992
Method:                 Least Squares   F-statistic:                     2628.
Date:                Mon, 27 Feb 2017   Prob (F-statistic):           7.75e-22
Time:                        19:57:29   Log-Likelihood:                -48.575
No. Observations:                  20   AIC:                             99.15
Df Residuals:                      19   BIC:                             100.1
Df Model:                           1                                         
Covariance Type:            nonrobust                                         
==============================================================================
                 coef    std err          t      P>|t|      [95.0% Conf. Int.]
------------------------------------------------------------------------------
x              0.0001   2.37e-06     51.269      0.000         0.000     0.000
==============================================================================
Omnibus:                        3.436   Durbin-Watson:                   1.028
Prob(Omnibus):                  0.179   Jarque-Bera (JB):                2.013
Skew:                           0.542   Prob(JB):                        0.366
Kurtosis:                       1.887   Cond. No.                         1.00
==============================================================================

Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

In [8]:

    

mod2 = smf.ols(formula='y_efficient ~ x - 1', data=d2).fit()
mod2.summary()


    

    
Out[8]:





OLS Regression Results

  
Dep. Variable:
       
y_efficient
   
  R-squared:         
 
   0.991


  
Model:
                   
OLS
       
  Adj. R-squared:    
 
   0.991


  
Method:
             
Least Squares
  
  F-statistic:       
 
   2163.


  
Date:
             
Mon, 27 Feb 2017
 
  Prob (F-statistic):
 
4.87e-21


  
Time:
                 
19:57:29
     
  Log-Likelihood:    
 
 -34.300


  
No. Observations:
      
    20
      
  AIC:               
 
   70.60


  
Df Residuals:
          
    19
      
  BIC:               
 
   71.59


  
Df Model:
              
     1
      
                     
     
 
   


  
Covariance Type:
      
nonrobust
    
                     
     
 
   




  
     
coef
     
std err
      
t
      
P>|t|
 
[95.0% Conf. Int.]
 


  
x
 
 5.388e-05
 
 1.16e-06
 
   46.510
 
 0.000
 
 5.15e-05  5.63e-05




  
Omnibus:
       
21.719
 
  Durbin-Watson:     
 
   2.000


  
Prob(Omnibus):
 
 0.000
 
  Jarque-Bera (JB):  
 
  28.717


  
Skew:
          
 1.952
 
  Prob(JB):          
 
5.81e-07


  
Kurtosis:
      
 7.384
 
  Cond. No.          
 
    1.00

In [9]:

    

y_est_basic = mod1.predict(x)
y_est_effi = mod2.predict(x)


    

In [12]:

    

fig, ax = plt.subplots(figsize=(10,10))
ax.scatter(x, y1, s=15)
ax.scatter(x, y2, s=15, c='g')
ax.plot(x, y_est_basic)
ax.plot(x, y_est_effi)
ax.legend(scatterpoints=1)


    

    
Out[12]:





<matplotlib.legend.Legend at 0x1029ac610>






    

In [ ]: