notebook.community

Edit and run



In [2]:

    
%matplotlib inline
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns



In [37]:

    
df = pd.read_csv('./combined_test_results.csv', index_col=0, header=0)



In [43]:

    
df.columns = ['Group', 'Threshold_Percentage', 
              'True_Positive', 'False_Positive', 
              'True_Negative', 'False_Negative', 
              'True_Positive_Rate', 'True_Negative_Rate', 
              'Positive_Predictive_Value', 'Negative_Predictive_Value', 
              'False_Negative_Rate', 'False_Positive_Rate',
              'False_Discovery_Rate', 'False_Omission_Rate', 
              'Accuracy', 'F1',
              'Matthews_Correlation_Coefficient', 'AUROC',
              'AUPR_0', 'AUPR'
             ]
df









    Out[43]:






  
    
      
      Group
      Threshold_Percentage
      True_Positive
      False_Positive
      True_Negative
      False_Negative
      True_Positive_Rate
      True_Negative_Rate
      Positive_Predictive_Value
      Negative_Predictive_Value
      False_Negative_Rate
      False_Positive_Rate
      False_Discovery_Rate
      False_Omission_Rate
      Accuracy
      F1
      Matthews_Correlation_Coefficient
      AUROC
      AUPR_0
      AUPR
    
  
  
    
      0
      error
      0
      304913
      37440
      476369
      52813
      0.85
      0.93
      0.89
      0.9000
      0.1500
      0.07
      0.11
      0.10
      0.90
      0.44
      0.79
      0.95
      0.965
      0.933
    
    
      1
      converged
      0
      136911
      27375
      665535
      41714
      0.77
      0.96
      0.83
      0.9400
      0.2300
      0.04
      0.17
      0.06
      0.92
      0.40
      0.75
      0.95
      0.987
      0.873
    
    
      2
      overall
      0
      13
      117
      870536
      869
      0.01
      1.00
      0.10
      1.0000
      0.9900
      0.00
      0.90
      0.00
      1.00
      0.01
      0.04
      0.55
      1.000
      0.019
    
    
      3
      overall
      25
      1122
      742
      865070
      4601
      0.20
      1.00
      0.60
      0.9900
      0.8000
      0.00
      0.40
      0.01
      0.99
      0.15
      0.34
      0.85
      0.999
      0.359
    
    
      4
      overall
      50
      3738
      1513
      858028
      8256
      0.31
      1.00
      0.71
      0.9900
      0.6900
      0.00
      0.29
      0.01
      0.99
      0.22
      0.47
      0.91
      0.999
      0.531
    
    
      5
      overall
      100
      10909
      3451
      844127
      13048
      0.46
      1.00
      0.76
      0.9800
      0.5400
      0.00
      0.24
      0.02
      0.98
      0.28
      0.58
      0.94
      0.998
      0.659
    
    
      6
      np
      0
      922
      1409
      863222
      5982
      0.13
      1.00
      0.40
      0.9931
      0.8665
      0.00
      0.60
      0.01
      0.99
      0.10
      0.23
      0.79
      0.998
      0.206
    
    
      7
      np
      25
      7200
      3627
      847004
      13704
      0.34
      1.00
      0.67
      0.9800
      0.6600
      0.00
      0.33
      0.02
      0.98
      0.23
      0.47
      0.91
      0.998
      0.526
    
    
      8
      np
      50
      13751
      5111
      835723
      16950
      0.45
      0.99
      0.73
      0.9800
      0.5500
      0.01
      0.27
      0.02
      0.97
      0.28
      0.56
      0.93
      0.997
      0.616
    
    
      9
      np
      100
      22438
      7113
      819845
      22139
      0.50
      0.99
      0.76
      0.9700
      0.5000
      0.01
      0.24
      0.03
      0.97
      0.30
      0.60
      0.94
      0.996
      0.682
    
    
      10
      sys
      0
      157
      464
      867456
      3458
      0.04
      1.00
      0.25
      1.0000
      0.9600
      0.00
      0.75
      0.00
      1.00
      0.04
      0.10
      0.60
      0.998
      0.055
    
    
      11
      sys
      25
      3762
      1727
      856098
      9948
      0.27
      1.00
      0.69
      0.9900
      0.7300
      0.00
      0.31
      0.01
      0.99
      0.20
      0.43
      0.88
      0.998
      0.465
    
    
      12
      sys
      50
      8994
      3106
      845950
      13485
      0.40
      1.00
      0.74
      0.9800
      0.6000
      0.00
      0.26
      0.02
      0.98
      0.26
      0.54
      0.92
      0.998
      0.608
    
    
      13
      sys
      100
      18136
      5332
      830213
      17854
      0.50
      0.99
      0.77
      0.9800
      0.5000
      0.01
      0.23
      0.02
      0.97
      0.31
      0.61
      0.94
      0.997
      0.693
    
    
      14
      np_and_system
      0
      8250
      3982
      846455
      12848
      0.39
      1.00
      0.67
      0.9900
      0.6100
      0.00
      0.33
      0.01
      0.98
      0.25
      0.50
      0.89
      0.997
      0.506
    
    
      15
      np_and_system
      25
      18085
      5671
      828924
      18855
      0.49
      0.99
      0.76
      0.9800
      0.5100
      0.01
      0.24
      0.02
      0.97
      0.30
      0.60
      0.94
      0.997
      0.675
    
    
      16
      np_and_system
      50
      24002
      7052
      819047
      21434
      0.53
      0.99
      0.77
      0.9700
      0.4700
      0.01
      0.23
      0.03
      0.97
      0.31
      0.62
      0.94
      0.997
      0.700
    
    
      17
      np_and_system
      100
      33235
      9699
      801718
      26883
      0.55
      0.99
      0.77
      0.9700
      0.4500
      0.01
      0.23
      0.03
      0.96
      0.32
      0.63
      0.95
      0.995
      0.723



In [44]:

    
labels=df.columns.values[2:]
labels









    Out[44]:





array(['True_Positive', 'False_Positive', 'True_Negative',
       'False_Negative', 'True_Positive_Rate', 'True_Negative_Rate',
       'Positive_Predictive_Value', 'Negative_Predictive_Value',
       'False_Negative_Rate', 'False_Positive_Rate',
       'False_Discovery_Rate', 'False_Omission_Rate', 'Accuracy', 'F1',
       'Matthews_Correlation_Coefficient', 'AUROC', 'AUPR_0', 'AUPR'], dtype=object)



In [40]:

    
# Compare AUROC results for all 18 labels
#     based on their groupings
label = 'AUROC'
sns.set_context("paper", font_scale=1.6)
sns.set_palette("Greys_r")
auroc_ax = sns.barplot(x='Group', y=label, 
                       hue='Threshold_Percentage', data=df)
auroc_ax.set(ylabel=label)
auroc_ax.legend(ncol=2, bbox_to_anchor=(1.05, 1),loc=0, frameon=True, title="Threshold Percentage")
plt.xticks(rotation=15)









    Out[40]:





(array([0, 1, 2, 3, 4, 5]), <a list of 6 Text xticklabel objects>)



In [41]:

    
# Compare TPR results for all 18 labels
#     based on their groupings
label = 'True_Positive_Rate'
sns.set_context("paper", font_scale=2)
sns.set_palette("Greys_r")
plt.figure(figsize=(12,10))
auroc_ax = sns.barplot(x='Group', y=label, 
                       hue='Threshold_Percentage', data=df)
auroc_ax.set(ylabel=label)
auroc_ax.set(xlabel="Classifier Group")
auroc_ax.legend(ncol=2, loc=0, frameon=True, title="Threshold Percentage")
plt.xticks(rotation=15)









    Out[41]:





(array([0, 1, 2, 3, 4, 5]), <a list of 6 Text xticklabel objects>)



In [46]:

    
# Compare TPR results for all 18 labels
#     based on their groupings
for lbl in labels:
    label = lbl
    sns.set_context("paper", font_scale=2.5)
    sns.set_palette("Greys_r", 6)
    plt.figure(figsize=(14,8))
    auroc_ax = sns.barplot(x='Group', y=label, 
                           hue='Threshold_Percentage', data=df)
    auroc_ax.set(ylabel=label)
    auroc_ax.set(xlabel="Classifier Group")
    auroc_ax.legend(ncol=2, bbox_to_anchor=(1.05, 1),loc=0, frameon=True, title="Threshold Percentage")
    #lgd = auroc_ax.legend(ncol=2, bbox_to_anchor=(1.05, 1),loc=0, frameon=True, title="Threshold Percentage")
    lgd = auroc_ax.legend(ncol=2, loc=0,frameon=True, title="Threshold Percentage")
    plt.xticks(rotation=15)
    plt.tight_layout()
    
    plt.savefig(label + '.png', bbox_extra_artists=(lgd,), bbox_inches='tight')
    plt.show()



In [ ]:

	Group	Threshold_Percentage	True_Positive	False_Positive	True_Negative	False_Negative	True_Positive_Rate	True_Negative_Rate	Positive_Predictive_Value	Negative_Predictive_Value	False_Negative_Rate	False_Positive_Rate	False_Discovery_Rate	False_Omission_Rate	Accuracy	F1	Matthews_Correlation_Coefficient	AUROC	AUPR_0	AUPR
0	error	0	304913	37440	476369	52813	0.85	0.93	0.89	0.9000	0.1500	0.07	0.11	0.10	0.90	0.44	0.79	0.95	0.965	0.933
1	converged	0	136911	27375	665535	41714	0.77	0.96	0.83	0.9400	0.2300	0.04	0.17	0.06	0.92	0.40	0.75	0.95	0.987	0.873
2	overall	0	13	117	870536	869	0.01	1.00	0.10	1.0000	0.9900	0.00	0.90	0.00	1.00	0.01	0.04	0.55	1.000	0.019
3	overall	25	1122	742	865070	4601	0.20	1.00	0.60	0.9900	0.8000	0.00	0.40	0.01	0.99	0.15	0.34	0.85	0.999	0.359
4	overall	50	3738	1513	858028	8256	0.31	1.00	0.71	0.9900	0.6900	0.00	0.29	0.01	0.99	0.22	0.47	0.91	0.999	0.531
5	overall	100	10909	3451	844127	13048	0.46	1.00	0.76	0.9800	0.5400	0.00	0.24	0.02	0.98	0.28	0.58	0.94	0.998	0.659
6	np	0	922	1409	863222	5982	0.13	1.00	0.40	0.9931	0.8665	0.00	0.60	0.01	0.99	0.10	0.23	0.79	0.998	0.206
7	np	25	7200	3627	847004	13704	0.34	1.00	0.67	0.9800	0.6600	0.00	0.33	0.02	0.98	0.23	0.47	0.91	0.998	0.526
8	np	50	13751	5111	835723	16950	0.45	0.99	0.73	0.9800	0.5500	0.01	0.27	0.02	0.97	0.28	0.56	0.93	0.997	0.616
9	np	100	22438	7113	819845	22139	0.50	0.99	0.76	0.9700	0.5000	0.01	0.24	0.03	0.97	0.30	0.60	0.94	0.996	0.682
10	sys	0	157	464	867456	3458	0.04	1.00	0.25	1.0000	0.9600	0.00	0.75	0.00	1.00	0.04	0.10	0.60	0.998	0.055
11	sys	25	3762	1727	856098	9948	0.27	1.00	0.69	0.9900	0.7300	0.00	0.31	0.01	0.99	0.20	0.43	0.88	0.998	0.465
12	sys	50	8994	3106	845950	13485	0.40	1.00	0.74	0.9800	0.6000	0.00	0.26	0.02	0.98	0.26	0.54	0.92	0.998	0.608
13	sys	100	18136	5332	830213	17854	0.50	0.99	0.77	0.9800	0.5000	0.01	0.23	0.02	0.97	0.31	0.61	0.94	0.997	0.693
14	np_and_system	0	8250	3982	846455	12848	0.39	1.00	0.67	0.9900	0.6100	0.00	0.33	0.01	0.98	0.25	0.50	0.89	0.997	0.506
15	np_and_system	25	18085	5671	828924	18855	0.49	0.99	0.76	0.9800	0.5100	0.01	0.24	0.02	0.97	0.30	0.60	0.94	0.997	0.675
16	np_and_system	50	24002	7052	819047	21434	0.53	0.99	0.77	0.9700	0.4700	0.01	0.23	0.03	0.97	0.31	0.62	0.94	0.997	0.700
17	np_and_system	100	33235	9699	801718	26883	0.55	0.99	0.77	0.9700	0.4500	0.01	0.23	0.03	0.96	0.32	0.63	0.95	0.995	0.723