

In [1]:

    
%matplotlib inline
import os, sys
import ntpath
from eden.util import display
# -----------------------------------------------------------------------------



In [2]:

    
from graph_finder import GraphFinder

Experiments:

read a directory of '.sto' files and create their graphs



In [3]:

    
pos_class_0_path = "StoExamples/StoClasses/pos_class_0"
pos_class_1_path = "StoExamples/StoClasses/pos_class_1"
neg_class_0_path = "StoExamples/StoClasses/neg_class_0"
neg_class_1_path = "StoExamples/StoClasses/neg_class_1"

pos_class_0_abs_path = os.path.abspath(pos_class_0_path)
pos_class_1_abs_path = os.path.abspath(pos_class_1_path)
neg_class_0_abs_path = os.path.abspath(neg_class_0_path)
neg_class_1_abs_path = os.path.abspath(neg_class_1_path)

convert files to graphs



In [ ]:

    
gf = GraphFinder()
pos_0_graphs = gf.convert(pos_class_0_abs_path)
pos_1_graphs = gf.convert(pos_class_1_abs_path)
neg_0_graphs = gf.convert(neg_class_0_abs_path)
neg_1_graphs = gf.convert(neg_class_1_abs_path)









    



{'graph_title': '550-70730-0-0'}
{'graph_title': '550-70852-0-0'}
{'graph_title': '550-53949-1-0'}
{'graph_title': '550-69275-0-0'}
{'graph_title': '550-69410-0-0'}
{'graph_title': '550-751-0-0'}
{'graph_title': '550-1137-0-0'}
{'graph_title': '550-904-1-0'}
{'graph_title': '550-1143-1-0'}
{'graph_title': '550-1153-0-0'}

draw the generated graphs



In [ ]:

    
for G in pos_0_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')
for G in pos_1_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')

for G in neg_0_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')
for G in neg_1_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')

transform the created graph to have the desired type of info(end user specification)



In [ ]:

    
transform_pos_0_graphs = gf.graphs_transform(pos_0_graphs, use_seq=True, use_cov = False)
transform_pos_1_graphs = gf.graphs_transform(pos_1_graphs, use_seq=True, use_cov = False)

transform_neg_0_graphs = gf.graphs_transform(neg_0_graphs, use_seq=True, use_cov = False)
transform_neg_1_graphs = gf.graphs_transform(neg_1_graphs, use_seq=True, use_cov = False)

draw the transformed graphs



In [ ]:

    
for G in pos_0_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')
for G in pos_1_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')

for G in neg_0_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')
for G in neg_1_graphs:
    display.draw_graph(G, size=40, node_size=400, font_size=20, node_border=True, prog='neato')

Classification

This Classification Class does:

Create target file.
Do the modeling.

Conversion

create target list



In [ ]:

    
import numpy as np

pos_0_neg_0 = [1]* len(pos_0_graphs) + [-1] * len(neg_0_graphs)
y_0_0 = np.asarray([int(value) for value in pos_0_neg_0])
pos_0_neg_1 = [1]* len(pos_0_graphs) + [-1] * len(neg_1_graphs)
y_0_1 = np.asarray([int(value) for value in pos_0_neg_1])

pos_1_neg_0 = [1]* len(pos_1_graphs) + [-1] * len(neg_0_graphs)
y_1_0 = np.asarray([int(value) for value in pos_1_neg_0])

pos_1_neg_1 = [1]* len(pos_1_graphs) + [-1] * len(neg_1_graphs)
y_1_1 = np.asarray([int(value) for value in pos_1_neg_1])

print y_0_0, y_0_1, y_1_0, y_1_1

load data and convert it to graphs



In [ ]:

    
from eden.converter.graph.gspan import gspan_to_eden

pos_0_neg_0_graphs = pos_0_graphs + neg_0_graphs
pos_0_neg_1_graphs = pos_0_graphs + neg_1_graphs
pos_1_neg_0_graphs = pos_1_graphs + neg_0_graphs
pos_1_neg_1_graphs = pos_1_graphs + neg_1_graphs

Vectorizer

'''setup the vectorizer'''



In [ ]:

    
from eden.graph import Vectorizer
vectorizer = Vectorizer(complexity=3)

extract features and build data matrix



In [ ]:

    
%%time
X_p0 = vectorizer.transform(pos_0_graphs)
print ('X_p0', X_p0)
print 'Instances: %d Features: %d with an avg of %d features per instance' % (X_p0.shape[0], X_p0.shape[1],  X_p0.getnnz()/X_p0.shape[0])

X_p1 = vectorizer.transform(pos_1_graphs)
print ('X_p1', X_p1)
print 'Instances: %d Features: %d with an avg of %d features per instance' % (X_p1.shape[0], X_p1.shape[1],  X_p1.getnnz()/X_p1.shape[0])

X_n0 = vectorizer.transform(neg_0_graphs)
print ('X_n0', X_n0)
print 'Instances: %d Features: %d with an avg of %d features per instance' % (X_n0.shape[0], X_n0.shape[1],  X_n0.getnnz()/X_n0.shape[0])

X_n1 = vectorizer.transform(neg_1_graphs)
print ('X_n1', X_n1)
print 'Instances: %d Features: %d with an avg of %d features per instance' % (X_n1.shape[0], X_n1.shape[1],  X_n1.getnnz()/X_n1.shape[0])

Modelling

the train data matrix



In [ ]:

    
from scipy.sparse import vstack
data_matrix_00 = vstack( [X_p0, X_n0], format="csr")
data_matrix_01 = vstack( [X_p0, X_n1], format="csr")
data_matrix_10 = vstack( [X_p1, X_n0], format="csr")
data_matrix_11 = vstack( [X_p1, X_n1], format="csr")

Induce a predictor and evaluate its performance



In [ ]:

    
import numpy as np
from sklearn.metrics import roc_auc_score
'''I know this is wrong. they should be from the same type, right!'''
score = roc_auc_score(y_1_1, data_matrix_00)



In [ ]: