In [1]:

    

# import required libraries to munge data
import pandas as pd
import numpy as np


    

In [2]:

    

# import matplotlib to draw plot inside of notebook
%matplotlib inline
from mpl_toolkits.mplot3d.axes3d import Axes3D
import matplotlib.pyplot as plt
from matplotlib import cm


    

In [3]:

    

%matplotlib inline
import acc_mgr as acc


    

In [ ]:

    

observed_iterations = [6400, 12800, 199680]    
observed_test_data = [256, 512, 2048, 8092]

for i in observed_iterations:
    for j in observed_test_data:
        fig = acc.plot_by_fc(i, j, to_log_scale=False)
        fig2 = acc.plot_by_fc(i, j, to_log_scale=True)
        file_name = 'acc_trend_' + str(i) + '_' + str(j)
        fig.savefig(file_name + '.svg', format='svg', dpi=900)
        file_name = 'acc_trend_' + str(i) + '_' + str(j) + "_log_scale"
        fig2.savefig(file_name + '.svg', format='svg', dpi=900)


    

In [ ]:

    

observed_iterations = [6400, 12800, 199680]
for i in observed_iterations:
    acc.plot_by_fc(i, 512, to_log_scale=False)


    

In [5]:

    

tables = acc.get_acc()
print len(tables)
tables.describe()


    

    




744






    
Out[5]:






  

    

      

      
Iteration
      
L1
      
L2
      
L3
      
Testing Accuracy(256)
      
Testing Accuracy(512)
      
Testing Accuracy(2048)
      
Testing Accuracy(8092)
    
  
  

    

      
count
      
744.000000
      
744.000000
      
744.000000
      
744.000000
      
744.000000
      
744.000000
      
744.000000
      
744.000000
    
    

      
mean
      
72960.000000
      
36.741935
      
113.322581
      
480.000000
      
0.553514
      
0.543708
      
0.531833
      
0.548819
    
    

      
std
      
89702.961031
      
42.688998
      
162.454899
      
343.392622
      
0.299698
      
0.291934
      
0.286335
      
0.294267
    
    

      
min
      
6400.000000
      
2.000000
      
2.000000
      
128.000000
      
0.058594
      
0.076170
      
0.083010
      
0.086510
    
    

      
25%
      
6400.000000
      
4.000000
      
8.000000
      
224.000000
      
0.249024
      
0.248050
      
0.239865
      
0.243557
    
    

      
50%
      
12800.000000
      
16.000000
      
32.000000
      
384.000000
      
0.589844
      
0.578120
      
0.563235
      
0.590335
    
    

      
75%
      
199680.000000
      
64.000000
      
128.000000
      
640.000000
      
0.835938
      
0.822758
      
0.800412
      
0.830573
    
    

      
max
      
199680.000000
      
128.000000
      
512.000000
      
1024.000000
      
0.996094
      
0.980470
      
0.967290
      
0.972810
    
  

In [ ]:

    




    

In [ ]:

    

print """
Findings:
  -  
"""