In [19]:

    

import pickle
%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np


    

In [11]:

    

df = pickle.load( open( "2015-12-12-mlpexperiments_results6.p", "rb" ) )
df.shape


    

    
Out[11]:





(50, 20)

In [12]:

    

df.head()


    

    
Out[12]:






  

    

      

      
cm_overall
      
epochs
      
network
      
pct_white
      
test_size
      
training_size
      
p_cm
      
P_cm
      
r_cm
      
R_cm
      
n_cm
      
N_cm
      
b_cm
      
B_cm
      
q_cm
      
Q_cm
      
k_cm
      
K_cm
      
white_cm
      
black_cm
    
  
  

    

      
1
      
[[3157, 2091], [1815, 2937]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[357, 246], [59, 511]]
      
[[519, 117], [329, 294]]
      
[[179, 248], [121, 324]]
      
[[319, 159], [269, 255]]
      
[[150, 155], [93, 223]]
      
[[326, 77], [254, 140]]
      
[[152, 154], [98, 204]]
      
[[313, 74], [274, 118]]
      
[[155, 213], [51, 298]]
      
[[325, 91], [289, 181]]
      
[[126, 162], [86, 198]]
      
[[236, 119], [168, 191]]
      
[[2038, 637], [1583, 1179]]
      
[[1119, 1178], [508, 1758]]
    
    

      
1
      
[[3030, 1726], [1942, 3302]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[440, 163], [151, 419]]
      
[[354, 282], [93, 530]]
      
[[232, 195], [205, 240]]
      
[[298, 180], [215, 309]]
      
[[195, 110], [132, 184]]
      
[[239, 164], [110, 284]]
      
[[191, 115], [135, 167]]
      
[[235, 152], [133, 259]]
      
[[227, 141], [113, 236]]
      
[[232, 184], [143, 327]]
      
[[167, 121], [134, 150]]
      
[[220, 135], [162, 197]]
      
[[1578, 1097], [856, 1906]]
      
[[1452, 845], [870, 1396]]
    
    

      
1
      
[[4131, 3462], [841, 1566]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[581, 22], [492, 78]]
      
[[511, 125], [270, 353]]
      
[[371, 56], [383, 62]]
      
[[317, 161], [282, 242]]
      
[[296, 9], [302, 14]]
      
[[327, 76], [237, 157]]
      
[[285, 21], [285, 17]]
      
[[295, 92], [244, 148]]
      
[[347, 21], [318, 31]]
      
[[315, 101], [248, 222]]
      
[[238, 50], [225, 59]]
      
[[248, 107], [176, 183]]
      
[[2013, 662], [1457, 1305]]
      
[[2118, 179], [2005, 261]]
    
    

      
1
      
[[3276, 2124], [1696, 2904]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[521, 82], [321, 249]]
      
[[353, 283], [74, 549]]
      
[[316, 111], [314, 131]]
      
[[242, 236], [164, 360]]
      
[[260, 45], [233, 83]]
      
[[224, 179], [98, 296]]
      
[[247, 59], [233, 69]]
      
[[211, 176], [104, 288]]
      
[[289, 79], [205, 144]]
      
[[203, 213], [101, 369]]
      
[[211, 77], [174, 110]]
      
[[199, 156], [103, 256]]
      
[[1432, 1243], [644, 2118]]
      
[[1844, 453], [1480, 786]]
    
    

      
1
      
[[2494, 1465], [2478, 3563]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[161, 442], [18, 552]]
      
[[483, 153], [230, 393]]
      
[[120, 307], [98, 347]]
      
[[287, 191], [237, 287]]
      
[[74, 231], [17, 299]]
      
[[290, 113], [201, 193]]
      
[[90, 216], [33, 269]]
      
[[277, 110], [220, 172]]
      
[[118, 250], [36, 313]]
      
[[280, 136], [179, 291]]
      
[[94, 194], [60, 224]]
      
[[220, 135], [136, 223]]
      
[[1837, 838], [1203, 1559]]
      
[[657, 1640], [262, 2004]]
    
  

In [13]:

    

#methods to decode serialized network json
import json

def get_num_layers(json_str):
    # number of actual layers - 5 for input and output / 3 for each hidden + 2 for input and output
    return (len(json.loads(json_str)['layers']) - 5) / 3 + 2

def get_first_activation(json_str):
    return json.loads(json_str)['layers'][1]['activation']

def get_first_width(json_str):
    return json.loads(json_str)['layers'][0]['output_dim']


    

In [14]:

    

#convert confusion matrices to accuracy
def cm2accuracy(cm):
    return (cm[0][0] + cm[1][1] * 1.0) / sum([sum(a) for a in cm])
    
df['overall_acc'] = df['cm_overall'].apply(cm2accuracy)
df['width'] = df['network'].apply(get_first_width)
df['num_layers'] = df['network'].apply(get_num_layers)


    

In [15]:

    

df.head()


    

    
Out[15]:






  

    

      

      
cm_overall
      
epochs
      
network
      
pct_white
      
test_size
      
training_size
      
p_cm
      
P_cm
      
r_cm
      
R_cm
      
...
      
B_cm
      
q_cm
      
Q_cm
      
k_cm
      
K_cm
      
white_cm
      
black_cm
      
overall_acc
      
width
      
num_layers
    
  
  

    

      
1
      
[[3157, 2091], [1815, 2937]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[357, 246], [59, 511]]
      
[[519, 117], [329, 294]]
      
[[179, 248], [121, 324]]
      
[[319, 159], [269, 255]]
      
...
      
[[313, 74], [274, 118]]
      
[[155, 213], [51, 298]]
      
[[325, 91], [289, 181]]
      
[[126, 162], [86, 198]]
      
[[236, 119], [168, 191]]
      
[[2038, 637], [1583, 1179]]
      
[[1119, 1178], [508, 1758]]
      
0.6094
      
512
      
3
    
    

      
1
      
[[3030, 1726], [1942, 3302]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[440, 163], [151, 419]]
      
[[354, 282], [93, 530]]
      
[[232, 195], [205, 240]]
      
[[298, 180], [215, 309]]
      
...
      
[[235, 152], [133, 259]]
      
[[227, 141], [113, 236]]
      
[[232, 184], [143, 327]]
      
[[167, 121], [134, 150]]
      
[[220, 135], [162, 197]]
      
[[1578, 1097], [856, 1906]]
      
[[1452, 845], [870, 1396]]
      
0.6332
      
512
      
3
    
    

      
1
      
[[4131, 3462], [841, 1566]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[581, 22], [492, 78]]
      
[[511, 125], [270, 353]]
      
[[371, 56], [383, 62]]
      
[[317, 161], [282, 242]]
      
...
      
[[295, 92], [244, 148]]
      
[[347, 21], [318, 31]]
      
[[315, 101], [248, 222]]
      
[[238, 50], [225, 59]]
      
[[248, 107], [176, 183]]
      
[[2013, 662], [1457, 1305]]
      
[[2118, 179], [2005, 261]]
      
0.5697
      
512
      
3
    
    

      
1
      
[[3276, 2124], [1696, 2904]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[521, 82], [321, 249]]
      
[[353, 283], [74, 549]]
      
[[316, 111], [314, 131]]
      
[[242, 236], [164, 360]]
      
...
      
[[211, 176], [104, 288]]
      
[[289, 79], [205, 144]]
      
[[203, 213], [101, 369]]
      
[[211, 77], [174, 110]]
      
[[199, 156], [103, 256]]
      
[[1432, 1243], [644, 2118]]
      
[[1844, 453], [1480, 786]]
      
0.6180
      
512
      
3
    
    

      
1
      
[[2494, 1465], [2478, 3563]]
      
1
      
{"layers": [{"b_constraint": null, "name": "De...
      
0.5028
      
10000
      
100000
      
[[161, 442], [18, 552]]
      
[[483, 153], [230, 393]]
      
[[120, 307], [98, 347]]
      
[[287, 191], [237, 287]]
      
...
      
[[277, 110], [220, 172]]
      
[[118, 250], [36, 313]]
      
[[280, 136], [179, 291]]
      
[[94, 194], [60, 224]]
      
[[220, 135], [136, 223]]
      
[[1837, 838], [1203, 1559]]
      
[[657, 1640], [262, 2004]]
      
0.6057
      
512
      
3
    
  

5 rows × 23 columns

In [16]:

    

x = df['epochs']
y = df['overall_acc']
plt.scatter(x,y)
plt.xlabel('epochs')
plt.ylabel('accuracy')
plt.ylim((0.0,1.0))


    

    
Out[16]:





(0.0, 1.0)






    

In [24]:

    

acc_mean = df.groupby(['epochs']).apply(np.mean)['overall_acc']
acc_std = df.groupby(['epochs']).apply(np.std)['overall_acc']
print(acc_std)
print(acc_mean)


    

    




epochs
1     0.021005
2     0.027110
3     0.040957
4     0.031917
5     0.014725
6     0.008749
7     0.014714
8     0.033315
9     0.021018
10    0.015690
Name: overall_acc, dtype: float64
epochs
1     0.60720
2     0.62432
3     0.65072
4     0.64384
5     0.67554
6     0.67440
7     0.67650
8     0.67976
9     0.71100
10    0.70580
Name: overall_acc, dtype: float64

In [43]:

    

# plot it!
fig, ax = plt.subplots(1, figsize=(12,8))
ax.plot(list(acc_mean.index), acc_mean, lw=2, color='blue')
ax.fill_between(list(acc_mean.index), acc_mean+acc_std, acc_mean-acc_std, facecolor='blue', alpha=0.2)
ax.set_xlabel('Epochs', fontsize=20)
ax.set_ylabel("Accuracy", fontsize=20)
ax.set_title("Epoch Count vs. Accuracy", fontsize=30)
plt.setp(ax.get_xticklabels(), fontsize=15)
plt.setp(ax.get_yticklabels(), fontsize=15)


    

    
Out[43]:





[None, None, None, None, None, None, None, None, None, None]






    

In [ ]: