Random Direct Graph

In [1]:

    

%matplotlib inline
import random
import matplotlib.pyplot as plt
from dgraph import standardize
from lesson1 import averageClustering as ac
from lesson2 import countEdges
from lesson4 import topk
from lesson4 import betweenness
from lesson5 import eigenvector_centrality as ec
from dgraph import randomDirectBalancedGraph as rdbg
from dgraph import Page_Rank as pr
from dgraph import fill_incoming as fi
from dgraph import Independent_Cascade as ic
from dgraph import toUndirect as und


    

In [2]:

    

def independent_cascade(graph, seed, centrality):
    adopters, haters, steps = ic(g, seed)
    print 'Independent Cascade Model: TOP %d %s' % (len(seed), centrality)
    print '\tFinal Adopters:\t', len(adopters)
    print '\tFinal Haters:\t', len(haters)
    print '\t# Iterations:\t', steps
    return len(adopters)


    

In [3]:

    

NODES = 7115
edges = random.randint(75000, 125000)
p = 0.5 # probability
seed = 100


    

In [4]:

    

g = rdbg(NODES, p, edges)
print 'Edges %d\tAverage Clustering = %f' % (countEdges(g)*2,ac(und(g)))
fi(g) # Fill incoming edges


    

    




Edges 78886	Average Clustering = 0.003204

In [6]:

    

print 'Page Rank execution...'
pagerank, iterations, err = pr(g, alpha=1.0e-5, eps=1.0e-3)
print iterations, ' iterations. Error:', err
top_pr = [a for a,b in topk(pagerank, seed)]


    

    




Page Rank execution...
12  iterations. Error: 0.000413387934502

In [7]:

    

print 'Eigenvector Centrality...',
cscores, diffsum = ec(g)
top_eigenc = [a for a, b in topk(cscores, seed)]
print 'Done'


    

    




Eigenvector Centrality... Done

In [8]:

    

print 'Betweennes centrality...',
bet = betweenness(g)
top_bet = [a for a, b in topk(bet, seed)]
print 'Done'


    

    




Betweennes centrality... Done

In [9]:

    

seed = 100
max_pr_ad = independent_cascade(g, top_pr[:seed], 'Page Rank')
max_eigenc_ad = independent_cascade(g, top_eigenc[:seed], 'Eigenvector')
max_bet_ad = independent_cascade(g, top_bet[:seed], 'Betweenness')


    

    




Independent Cascade Model: TOP 100 Page Rank
	Final Adopters:	179
	Final Haters:	6936
	# Iterations:	7
Independent Cascade Model: TOP 100 Eigenvector
	Final Adopters:	267
	Final Haters:	6848
	# Iterations:	5
Independent Cascade Model: TOP 100 Betweenness
	Final Adopters:	201
	Final Haters:	6914
	# Iterations:	5

In [10]:

    

while seed > 0:
    seed -= 5
    adopters_pr = independent_cascade(g, top_pr[:seed], 'Page Rank')
    if max_pr_ad <= adopters_pr:
        max_pr_ad = adopters_pr
    else:
        break


    

    




Independent Cascade Model: TOP 95 Page Rank
	Final Adopters:	152
	Final Haters:	6963
	# Iterations:	5

In [11]:

    

seed = 100
while seed > 0:
    seed -= 5
    adopters_eigenc = independent_cascade(g, top_eigenc[:seed], 'Eigenvector')
    if max_eigenc_ad <= adopters_eigenc:
        max_eigenc_ad = adopters_eigenc
    else:
        break


    

    




Independent Cascade Model: TOP 95 Eigenvector
	Final Adopters:	254
	Final Haters:	6861
	# Iterations:	7

In [12]:

    

seed = 100
while seed > 0:
    seed -= 5
    adopters_bet = independent_cascade(g, top_bet[:seed], 'Betweenness')
    if max_bet_ad <= adopters_bet:
        max_bet_ad = adopters_bet
    else:       
        break


    

    




Independent Cascade Model: TOP 95 Betweenness
	Final Adopters:	226
	Final Haters:	6889
	# Iterations:	7
Independent Cascade Model: TOP 90 Betweenness
	Final Adopters:	223
	Final Haters:	6892
	# Iterations:	10

In [13]:

    

max_adopters = (max_pr_ad, max_eigenc_ad, max_bet_ad)

fig, ax = plt.subplots()

bar_width = 0.35

opacity = 0.4

rects1 = plt.bar(1, max_pr_ad, bar_width,
                 alpha=opacity,
                 color='b',
                 label='Page Rank')

rects2 = plt.bar(2, max_eigenc_ad, bar_width,
                 alpha=opacity,
                 color='r',
                 label='Eigenvector')

rects3 = plt.bar(3, max_bet_ad, bar_width,
                 alpha=opacity,
                 color='y',
                 label='Betweenness')

plt.xlabel('Centrality Measures')
plt.ylabel('Adopters')
plt.title('Adopters Comparison')
plt.xticks([1.2,2.2,3.2], ('PR', 'E', 'B'))

plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)

plt.tight_layout()