In [1]:

    

%cd ~/NetBeansProjects/ExpLosion/
%load_ext autoreload
from notebooks.common_imports import *
from gui.output_utils import *

sns.timeseries.algo.bootstrap = my_bootstrap
sns.categorical.bootstrap = my_bootstrap


    

    




/Volumes/LocalDataHD/m/mm/mmb28/NetBeansProjects/ExpLosion

In [2]:

    

s = {'document_features_ev': 'AN+NN',
     'document_features_tr': 'J+N+AN+NN',
     'expansions__decode_handler': 'SignifiedOnlyFeatureHandler',
     'expansions__entries_of_id': None,
     'expansions__k': 3,
     'expansions__neighbour_strategy': 'linear',
     'expansions__noise': 0.0,
     'expansions__use_random_neighbours': 0,
     'expansions__use_similarity': 0,
     'expansions__vectors__algorithm': 'word2vec',
#      'expansions__vectors__composer': 'Add',
     'expansions__vectors__dimensionality': 100,
     'expansions__vectors__rep': 0,
     'expansions__vectors__unlabelled': 'wiki',
     'expansions__vectors__unlabelled_percentage__in': [100],
     'labelled': 'amazon_grouped-tagged'}
ids = Experiment.objects.filter(**s).values_list('id', flat=True)
fields = {'unlab': 'expansions__vectors__unlabelled',
            'percent': 'expansions__vectors__unlabelled_percentage',
            'Composer': 'expansions__vectors__composer',
            'Overlap': 'expansions__allow_overlap'}
print('ids are', ids, 'total', len(ids))
df = dataframe_from_exp_ids(ids, fields)
df['corpus'] = ['%s-%s'%(a,b) for a,b in zip(df.unlab, df.percent)]
df.Overlap = df.Overlap.map({'1': 'Allowed', '0': 'Not allowed'})


    

    




ids are [34, 35, 36, 37, 209, 212, 215, 218] total 8
Calculating CI for exp 209
Calculating CI for exp 212
Calculating CI for exp 215
Calculating CI for exp 218
Overlap has 4000 values
unlab has 4000 values
folds has 4000 values
percent has 4000 values
Composer has 4000 values
Accuracy has 4000 values

In [3]:

    

sns.timeseries.algo.bootstrap = my_bootstrap
sns.categorical.bootstrap = my_bootstrap
with sns.color_palette("cubehelix", 4):
    g = sns.factorplot(y='Accuracy', hue='Composer', x='Overlap', 
                       x_order=['Not allowed', 'Allowed'], 
                       data=df, kind='bar', ci=68, aspect=1.5);
for ax in g.axes.flat:
    ax.axhline(random_vect_baseline(), c='k');
    ax.axhline(nondistributional_baseline(**settings_of(ids[0])), c='k');
g.savefig('plot-nps-with-lexical-overlap.pdf', format='pdf', dpi=300, bbox_inches='tight', pad_inches=0.1)


    

    




/home/m/mm/mmb28/anaconda3/lib/python3.4/site-packages/seaborn/categorical.py:2653: UserWarning: The `x_order` parameter has been renamed `order`
  UserWarning)

WARNING:py.warnings:/home/m/mm/mmb28/anaconda3/lib/python3.4/site-packages/seaborn/categorical.py:2653: UserWarning: The `x_order` parameter has been renamed `order`
  UserWarning)







    

In [4]:

    

df.groupby(['Overlap', 'Composer']).mean()


    

    
Out[4]:






  

    

      

      

      
Accuracy
      
folds
    
    

      
Overlap
      
Composer
      

      

    
  
  

    

      
Allowed
      
Add
      
0.829048
      
249.5
    
    

      
Left
      
0.678007
      
249.5
    
    

      
Mult
      
0.801022
      
249.5
    
    

      
Right
      
0.708641
      
249.5
    
    

      
Not allowed
      
Add
      
0.529362
      
249.5
    
    

      
Left
      
0.352687
      
249.5
    
    

      
Mult
      
0.471030
      
249.5
    
    

      
Right
      
0.343514
      
249.5
    
  

In [8]:

    

from gui.user_code import pairwise_significance_exp_ids
corpus='amazon_grouped-tagged'
document_features_tr='J+N+AN+NN'
document_features_ev='AN+NN'
res = Experiment.objects.get(labelled=corpus,
                             document_features_tr=document_features_tr,
                             document_features_ev=document_features_ev,
                             expansions__decode_handler='BaseFeatureHandler')
pairwise_significance_exp_ids(zip(ids, [res.id]*len(ids)))


    

    




Running significance for (34, 5)
Running significance for (35, 5)
Running significance for (36, 5)
Running significance for (37, 5)
Running significance for (209, 5)
Running significance for (212, 5)
Running significance for (215, 5)
Running significance for (218, 5)






    
Out[8]:






  

    

      

      
name1
      
acc1
      
name2
      
acc2
      
mean_diff
      
pval
      
significant
    
  
  

    

      
0
      
34
      
0.529542
      
5
      
0.845445
      
0.315903
      
0
      
True
    
    

      
1
      
35
      
0.470981
      
5
      
0.845445
      
0.374463
      
0
      
True
    
    

      
2
      
36
      
0.352811
      
5
      
0.845445
      
0.492634
      
0
      
True
    
    

      
3
      
37
      
0.343414
      
5
      
0.845445
      
0.502030
      
0
      
True
    
    

      
4
      
209
      
0.828949
      
5
      
0.845445
      
0.016496
      
0
      
True
    
    

      
5
      
212
      
0.800792
      
5
      
0.845445
      
0.044653
      
0
      
True
    
    

      
6
      
215
      
0.677613
      
5
      
0.845445
      
0.167832
      
0
      
True
    
    

      
7
      
218
      
0.708527
      
5
      
0.845445
      
0.136917
      
0
      
True
    
  

In [ ]: