-
Notifications
You must be signed in to change notification settings - Fork 0
/
testGraphDistance.py
65 lines (53 loc) · 1.94 KB
/
testGraphDistance.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
import sys
import random
import operator
import copy
from collections import OrderedDict
import math
import linkpred
if __name__== '__main__':
if sys.argv is None or len(sys.argv) is not 3:
print "Usage : python testRootedPageRank.py ddi_v2.0.txt ddi_v3.0.txt"
exit()
trainfile = file(sys.argv[1])
testfile= file(sys.argv[2])
trainnet = dict()
testnet = dict()
i=0
for line in trainfile:
line=line.strip().split("\t")
#if len(line) != 2: continue
node1=line[0]
node2=line[1]
# add link to the network (dictionary of node-neighborlist pair)
linkpred.addLinkToNetwork(trainnet,node1, node2)
linkpred.addLinkToNetwork(trainnet,node2, node1)
i+=1
for line in testfile:
line=line.strip().split("\t")
#if len(line) != 2: continue
node1=line[0]
node2=line[1]
linkpred.addLinkToNetwork(testnet,node1, node2)
linkpred.addLinkToNetwork(testnet,node2, node1)
#weights = linkpred.computeRankedList(wholenet, wholenet.keys(), {}, func=rootedPagerank)
#print weights
#ranks= linkpred.pagerank(wholenet,weights=weights)
#print sorted(ranks.items(),key=lambda x:x[1])
testnet = linkpred.getDifference(testnet,trainnet)
#core = set(trainnet.keys()).intersection(testnet)
core = linkpred.getCoreNodes(trainnet,testnet,1,1)
testnet_pruned = linkpred.prune(testnet,core)
nLinks=linkpred.getNumberOfLinks(testnet_pruned)
print "trainnet :", linkpred.getNumberOfLinks(trainnet)
print "testnet :",nLinks
#rankedList = linkpred.computeRankedList(trainnet, core, trainnet)
rankedList = linkpred.computeRankedList(trainnet, core, trainnet, func=linkpred.graphDistance)
ranking=OrderedDict(sorted(rankedList.items(),key=operator.itemgetter(1),reverse=True))
#print ranking
auc = linkpred.computeAUCBySampling(testnet_pruned,ranking)
intersect =linkpred.computeTruePositive(ranking, testnet_pruned, nLinks)
prec = intersect/float(nLinks)
print "Correctly predicted :", intersect
print "AUC :",auc
print "Precision :" , prec