def binarySearchBoundary(G, k, Tsize, targeted_size, step, p, iterations):
# initialization for binary search
R = iterations
stepk = -int(math.ceil(float(step)/2))
k += stepk
if k not in Tsize:
S = degreeDiscountIC(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T))/R
Tsize[k] = avg
# check values of Tsize in between last 2 calculated steps
while stepk != 1:
print k, stepk, Tsize[k]
if Tsize[k] >= targeted_size:
stepk = -int(math.ceil(float(abs(stepk))/2))
else:
stepk = int(math.ceil(float(abs(stepk))/2))
k += stepk
if k not in Tsize:
S = degreeDiscountIC(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T))/R
Tsize[k] = avg
return S, Tsize
def binarySearchBoundary(G, k, Tsize, targeted_size, step, p, iterations):
# initialization for binary search
R = iterations
stepk = -int(math.ceil(float(step) / 2))
k += stepk
if k not in Tsize:
S = degreeDiscountIC(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T)) / R
Tsize[k] = avg
# check values of Tsize in between last 2 calculated steps
while stepk != 1:
print k, stepk, Tsize[k]
if Tsize[k] >= targeted_size:
stepk = -int(math.ceil(float(abs(stepk)) / 2))
else:
stepk = int(math.ceil(float(abs(stepk)) / 2))
k += stepk
if k not in Tsize:
S = degreeDiscountIC(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T)) / R
Tsize[k] = avg
return S, Tsize
def getDDData(G, maxk, p):
data = dict()
for i in range(1, maxk + 1):
S = degreeDiscountIC(G, i, p)
size = avgSize(G, S, p, 200)
data[i] = size
return data
def spreadDegreeDiscount(G, targeted_size, step=1, p=.01, iterations=200):
''' Finds initial set of nodes to propagate in Independent Cascade model (with priority queue)
Input: G -- networkx graph object
targeted_size -- desired size of targeted set
step -- step after each to calculate spread
p -- propagation probability
R -- number of iterations to average influence spread
Output:
S -- seed set that achieves targeted_size
Tsize -- averaged targeted size for different sizes of seed set
'''
Tsize = dict()
k = 0
Tsize[k] = 0
R = iterations
while Tsize[k] <= targeted_size:
k += step
S = degreeDiscountIC(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T))/R
Tsize[k] = avg
print k, Tsize[k]
# binary search for optimal solution
return binarySearchBoundary(G, k, Tsize, targeted_size, step, p, iterations)
def spreadDegreeDiscount(G, targeted_size, step=1, p=.01, iterations=200):
''' Finds initial set of nodes to propagate in Independent Cascade model (with priority queue)
Input: G -- networkx graph object
targeted_size -- desired size of targeted set
step -- step after each to calculate spread
p -- propagation probability
R -- number of iterations to average influence spread
Output:
S -- seed set that achieves targeted_size
Tsize -- averaged targeted size for different sizes of seed set
'''
Tsize = dict()
k = 0
Tsize[k] = 0
R = iterations
while Tsize[k] <= targeted_size:
k += step
S = degreeDiscountIC(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T)) / R
Tsize[k] = avg
print k, Tsize[k]
# binary search for optimal solution
return binarySearchBoundary(G, k, Tsize, targeted_size, step, p,
iterations)
def spreadDegreeDiscount(G, targeted_size, step=1, p=.01, iterations=200, degreeDiscount_Heuristic='degreeDiscount'):
''' Finds initial set of nodes to propagate in Independent Cascade model
Input: G -- networkx graph object
targeted_size -- desired size of targeted set
step -- step after each to calculate spread
p -- propagation probability
R -- number of iterations to average influence spread
Output:
S -- seed set that achieves targeted_size
Tsize -- averaged targeted size for different sizes of seed set
'''
# calculate the time of selecting the first initial nodes (afterwards we will select some of them)
start_time = time.time()
Tsize = dict()
k = 0
Tsize[k] = 0
R = iterations
while Tsize[k] <= targeted_size:
k += step
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k, p)
else:
S = degreeHeuristic(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T))/R
Tsize[k] = avg
print(k, Tsize[k])
# binary search for optimal solution
return binarySearchBoundary(G, k, Tsize, targeted_size, step, p, iterations, initial_time=time.time()-start_time)
G = nx.Graph()
with open('graphdata/../graphdata/hep.txt') as f:
n, m = f.readline().split()
for line in f:
u, v = map(int, line.split())
try:
G[u][v]['weight'] += 1
except:
G.add_edge(u, v, weight=1)
# G.add_edge(u, v, weight=1)
print 'Built graph G'
print time.time() - start
#calculate initial set
seed_size = 10
S = degreeDiscountIC(G, seed_size)
print 'Initial set of', seed_size, 'nodes chosen'
print time.time() - start
# write results S to file
with open('visualisation.txt', 'w') as f:
for node in S:
f.write(str(node) + os.linesep)
# calculate average activated set size
iterations = 200 # number of iterations
avg = 0
for i in range(iterations):
T = runIC(G, S)
avg += float(len(T)) / iterations
# print i, 'iteration of IC'
def binarySearchBoundary(G, k, Tsize, targeted_size, step, p, iterations, degreeDiscount_Heuristic='degreeDiscount', initial_time=0):
# Calculate the time it takes to select each node, initial_time is added in order keep track of the time needed for
# the selection of initial nodes
start_time = time.time()
# keep a list for time needed to select each nodes
timer_each_node = []
timer_each_node.append(initial_time)
# initialization for binary search
R = iterations
stepk = -int(math.ceil(float(step)/2))
k += stepk
if k not in Tsize:
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k, p)
else:
S = degreeHeuristic(G, k, p)
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T))/R
timer_each_node.append(time.time() + initial_time - start_time)
Tsize[k] = avg
# check values of Tsize in between last 2 calculated steps
while stepk != 1:
print(k, stepk, Tsize[k])
if Tsize[k] >= targeted_size:
stepk = -int(math.ceil(float(abs(stepk))/2))
else:
stepk = int(math.ceil(float(abs(stepk))/2))
k += stepk
if k not in Tsize:
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k, p)
else:
S = degreeHeuristic(G, k, p)
# stores the influence spread of each node
influence_spread = []
avg = 0
for i in range(R):
T = runIC(G, S, p)
avg += float(len(T))/R
influence_spread.append(avg)
# keep time for each NEW node selected
timer_each_node.append(time.time() + initial_time - start_time)
Tsize[k] = avg
print("datafaq: ", Tsize)
print("datafaq[k]: ", Tsize[k])
print("leeeen: ", len(timer_each_node))
return S, influence_spread, timer_each_node
def binaryDegreeDiscount(G, tsize, p=.01, a=0.38, step=5, iterations=200):
''' Finds minimal number of nodes necessary to reach tsize number of nodes
using degreeDiscount algorithms and binary search.
Input: G -- networkx graph object
tsize -- number of nodes necessary to reach
p -- propagation probability
a -- fraction of tsize to use as initial seed set size
step -- step between iterations of binary search
iterations -- number of iterations to average independent cascade
Output:
S -- seed set
Tspread -- spread values for different sizes of seed set
'''
Tspread = dict()
# find initial total spread
k0 = int(a * tsize)
S = degreeDiscountIC(G, k0, p)
t = avgSize(G, S, p, iterations)
Tspread[k0] = t
# find bound (lower or upper) of total spread
k = k0
print k, step, Tspread[k]
if t >= tsize:
# find the value of k that doesn't spread influence up to tsize nodes
step *= -1
while t >= tsize:
# reduce step if necessary
while k + step < 0:
step = int(math.ceil(float(step) / 2))
k += step
S = degreeDiscountIC(G, k, p)
t = avgSize(G, S, p, iterations)
Tspread[k] = t
print k, step, Tspread[k]
else:
# find the value of k that spreads influence up to tsize nodes
while t < tsize:
k += step
S = degreeDiscountIC(G, k, p)
t = avgSize(G, S, p, iterations)
Tspread[k] = t
print k, step, Tspread[k]
if Tspread[k] < Tspread[k - step]:
k -= step
step = abs(step)
# search precise boundary
stepk = step
while abs(stepk) != 1:
if Tspread[k] >= tsize:
stepk = -int(math.ceil(float(abs(stepk)) / 2))
else:
stepk = int(math.ceil(float(abs(stepk)) / 2))
k += stepk
if k not in Tspread:
S = degreeDiscountIC(G, k, p)
Tspread[k] = avgSize(G, S, p, iterations)
print k, stepk, Tspread[k]
return S, Tspread
#S = randomHeuristic(G, seed_size, p=.05)
#S = newGreedyICRev(G, seed_size, I, pi, pa, p=.05)
#S = newGreedyIC(G, seed_size, p=.05)
time1 = time.clock() - start
iterations = 200 # number of iterations
avg1 = 0
avg2 = 0
for i in range(iterations):
T, In = runIC(G, S, I)
avg1 += float(len(T)) / iterations
avg2 += float(len(In)) / iterations
pr_res = str(seed_size)
l1 = len(S)
rev1 = int(round(avg1)) + pi * int(round(avg2)) - c1 * l1
c2 = 1.2 #seed node cost=1.2
S = degreeDiscountIC(G, seed_size, p=.05)
#S = Rev(G, seed_size, I, pi, pa, c2)
avg1 = 0
avg2 = 0
for i in range(iterations):
T, In = runIC(G, S, I)
avg1 += float(len(T)) / iterations
avg2 += float(len(In)) / iterations
l2 = len(S)
rev2 = int(round(avg1)) + pi * int(round(avg2)) - c2 * l2
pr_res += '\t' + str(round(rev1,3)) + '\t' + str(round(rev2,3)) + '\t' + str(l1) + '\t' + str(l2)+ '\t' + str(time1) + '\n'
print(pr_res)
def binaryDegreeDiscount(G, tsize, p=0.01, a=0.38, step=5, iterations=200):
""" Finds minimal number of nodes necessary to reach tsize number of nodes
using degreeDiscount algorithms and binary search.
Input: G -- networkx graph object
tsize -- number of nodes necessary to reach
p -- propagation probability
a -- fraction of tsize to use as initial seed set size
step -- step between iterations of binary search
iterations -- number of iterations to average independent cascade
Output:
S -- seed set
Tspread -- spread values for different sizes of seed set
"""
Tspread = dict()
# find initial total spread
k0 = int(a * tsize)
S = degreeDiscountIC(G, k0, p)
t = avgSize(G, S, p, iterations)
Tspread[k0] = t
# find bound (lower or upper) of total spread
k = k0
print k, step, Tspread[k]
if t >= tsize:
# find the value of k that doesn't spread influence up to tsize nodes
step *= -1
while t >= tsize:
# reduce step if necessary
while k + step < 0:
step = int(math.ceil(float(step) / 2))
k += step
S = degreeDiscountIC(G, k, p)
t = avgSize(G, S, p, iterations)
Tspread[k] = t
print k, step, Tspread[k]
else:
# find the value of k that spreads influence up to tsize nodes
while t < tsize:
k += step
S = degreeDiscountIC(G, k, p)
t = avgSize(G, S, p, iterations)
Tspread[k] = t
print k, step, Tspread[k]
if Tspread[k] < Tspread[k - step]:
k -= step
step = abs(step)
# search precise boundary
stepk = step
while abs(stepk) != 1:
if Tspread[k] >= tsize:
stepk = -int(math.ceil(float(abs(stepk)) / 2))
else:
stepk = int(math.ceil(float(abs(stepk)) / 2))
k += stepk
if k not in Tspread:
S = degreeDiscountIC(G, k, p)
Tspread[k] = avgSize(G, S, p, iterations)
print k, stepk, Tspread[k]
return S, Tspread
G = nx.Graph()
with open('graphdata/../graphdata/hep.txt') as f:
n, m = f.readline().split()
for line in f:
u, v = map(int, line.split())
try:
G[u][v]['weight'] += 1
except:
G.add_edge(u,v, weight=1)
# G.add_edge(u, v, weight=1)
print 'Built graph G'
print time.time() - start
#calculate initial set
seed_size = 10
S = degreeDiscountIC(G, seed_size)
print 'Initial set of', seed_size, 'nodes chosen'
print time.time() - start
# write results S to file
with open('visualisation.txt', 'w') as f:
for node in S:
f.write(str(node) + os.linesep)
# calculate average activated set size
iterations = 200 # number of iterations
avg = 0
for i in range(iterations):
T = runIC(G, S)
avg += float(len(T))/iterations
# print i, 'iteration of IC'
def binaryDegreeDiscount(G,
tsize,
p=.01,
a=0.38,
step=5,
iterations=200,
degreeDiscount_Heuristic='degreeDiscount'):
''' Finds minimal number of nodes necessary to reach tsize number of nodes
using degreeDiscount algorithms and binary search.
Input: G -- networkx graph object
tsize -- number of nodes necessary to reach
p -- propagation probability
a -- fraction of tsize to use as initial seed set size
step -- step between iterations of binary search
iterations -- number of iterations to average independent cascade
degreeDiscount_Heuristic -- whether to select degree Discount or degree Heuristic algorithm
Output:
S -- seed set
Tspread -- spread values for different sizes of seed set
'''
# Calculate the time it takes to select each node
start_time = time.time()
# keep a list for time needed to select each nodes
timer_each_node = []
Tspread = dict()
# find initial total spread
k0 = int(a * tsize)
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k0, p)
else:
S = degreeHeuristic(G, k0, p)
t = avgSize(G, S, p, iterations)
Tspread[k0] = t
# find bound (lower or upper) of total spread
k = k0
print(k, step, Tspread[k])
# keep time for each NEW node selected
timer_each_node.append(time.time() - start_time)
if t >= tsize:
# find the value of k that doesn't spread influence up to tsize nodes
step *= -1
while t >= tsize:
# reduce step if necessary
while k + step < 0:
step = int(math.ceil(float(step) / 2))
k += step
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k, p)
else:
S = degreeHeuristic(G, k, p)
t = avgSize(G, S, p, iterations)
Tspread[k] = t
print(k, step, Tspread[k])
# keep time for each NEW node selected
timer_each_node.append(time.time() - start_time)
else:
# find the value of k that spreads influence up to tsize nodes
while t < tsize:
k += step
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k, p)
else:
S = degreeHeuristic(G, k, p)
t = avgSize(G, S, p, iterations)
Tspread[k] = t
print(k, step, Tspread[k])
# keep time for each NEW node selected
timer_each_node.append(time.time() - start_time)
if Tspread[k] < Tspread[k - step]:
k -= step
step = abs(step)
# search precise boundary
stepk = step
while abs(stepk) != 1:
if Tspread[k] >= tsize:
stepk = -int(math.ceil(float(abs(stepk)) / 2))
else:
stepk = int(math.ceil(float(abs(stepk)) / 2))
k += stepk
if k not in Tspread:
if (degreeDiscount_Heuristic == 'degreeDiscount'):
S = degreeDiscountIC(G, k, p)
else:
S = degreeHeuristic(G, k, p)
Tspread[k] = avgSize(G, S, p, iterations)
# keep time for each NEW node selected
timer_each_node.append(time.time() - start_time)
print(k, stepk, Tspread[k])
print("(number of nodes) - (spread) : ", Tspread)
# ======================================================================================================================
# stores the influence spread of each node
influence_spread = []
avg = 0
for i in range(iterations):
T = runIC(G, S, p)
avg += float(len(T)) / iterations
influence_spread.append(avg)
print("(spread) : ", influence_spread)
return S, influence_spread, timer_each_node
try:
G[u][v]['weight'] += 1
except:
G.add_edge(u, v, weight=1)
# G.add_edge(u, v, weight=1)
print 'Built graph G'
print time.time() - start
seed_size = 5
p = .01
nodes = G.nodes()
C = combinations(nodes, seed_size)
spread = dict()
for candidate in C:
print candidate,
time2spread = time.time()
spread[candidate] = avgSize(G, list(candidate), p, 1000)
print spread[candidate], time.time() - time2spread
S, val = max(spread.iteritems(), key=lambda (dk, dv): dv)
print 'S (by brute-force):', S, ' -->', val
S2 = degreeDiscountIC(G, seed_size, p)
print 'S (by degree discount):', tuple(S2), ' -->', avgSize(G, S2, p, 1000)
print 'S (by degree discount) spreads to %s nodes (according to brute-force)' % (
spread[tuple(sorted(S2))])
print 'Total time:', time.time() - start
console = []
u, v = map(int, line.split())
try:
G[u][v]['weight'] += 1
except:
G.add_edge(u,v, weight=1)
# G.add_edge(u, v, weight=1)
print 'Built graph G'
print time.time() - start
seed_size = 5
p = .01
nodes = G.nodes()
C = combinations(nodes, seed_size)
spread = dict()
for candidate in C:
print candidate,
time2spread = time.time()
spread[candidate] = avgSize(G, list(candidate), p, 1000)
print spread[candidate], time.time() - time2spread
S, val = max(spread.iteritems(), key = lambda (dk, dv): dv)
print 'S (by brute-force):', S, ' -->', val
S2 = degreeDiscountIC(G, seed_size, p)
print 'S (by degree discount):', tuple(S2), ' -->', avgSize(G, S2, p, 1000)
print 'S (by degree discount) spreads to %s nodes (according to brute-force)' %(spread[tuple(sorted(S2))])
print 'Total time:', time.time() - start
console = []
