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
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
|
import matplotlib.pyplot as plt
import numpy as np
import cascade_creation
import convex_optimization
import algorithms
import rip_condition
def compare_greedy_and_lagrange_cs284r():
"""
Compares the performance of the greedy algorithm on the
lagrangian sparse recovery objective on the Facebook dataset
for the CS284r project
"""
G = cascade_creation.InfluenceGraph(max_proba = .8)
G.import_from_file("../datasets/subset_facebook_SNAPnormalize.txt")
A = cascade_creation.generate_cascades(G, p_init=.05, n_cascades=100)
#Greedy
G_hat = algorithms.greedy_prediction(G, A)
algorithms.correctness_measure(G, G_hat, print_values=True)
#Lagrange Objective
G_hat = algorithms.recovery_l1obj_l2constraint(G, A,
passed_function=convex_optimization.type_lasso,
floor_cstt=.05, lbda=10)
algorithms.correctness_measure(G, G_hat, print_values=True)
def compute_graph(graph_name, n_cascades, lbda, passed_function, min_proba,
max_proba, sparse_edges=False, p_init=.05):
"""
Test running time on different algorithms
"""
G = cascade_creation.InfluenceGraph(max_proba=max_proba,
min_proba=min_proba,
sparse_edges=sparse_edges)
G.import_from_file(graph_name)
A = cascade_creation.generate_cascades(G, p_init=p_init, n_cascades=n_cascades)
if passed_function==algorithms.greedy_prediction:
G_hat = algorithms.greedy_prediction(G, A)
else:
G_hat = algorithms.recovery_passed_function(G, A,
passed_function=passed_function,
floor_cstt=.1, lbda=lbda, n_cascades=n_cascades)
algorithms.correctness_measure(G, G_hat, print_values=True)
def plot_watts_strogatz_graph():
"""
plot information in a pretty way
"""
plt.clf()
fig = plt.figure(1)
labels = [50, 100, 500, 1000, 2000, 5000]
x = [np.log(50), np.log(100), np.log(500),
np.log(1000), np.log(2000), np.log(5000)]
sparse_recov = [.25, .32, .7, .82, .89, .92]
max_likel = [.21, .29, .67, .8, .87, .9]
lasso = [.07, .30, .46, .65, .86, .89]
greedy = [.09, .15, .4, .63, .82, .92]
fig, ax = plt.subplots()
plt.axis((np.log(45), np.log(5500), 0, 1))
plt.xlabel("Number of Cascades")
plt.ylabel("F1 score")
plt.grid(color="lightgrey")
ax.plot(x, greedy, 'ko-', color="lightseagreen", label='Greedy')
ax.plot(x, lasso, 'ko-', color="orange", label="Lasso")
ax.plot(x, max_likel, 'ko-', color="cornflowerblue", label="MLE")
ax.plot(x, sparse_recov, 'ko-', color="k", label="Our Method")
plt.legend(loc="lower right")
ax.set_xticks(x)
ax.set_xticklabels(tuple(labels))
plt.savefig("../paper/figures/"+"watts_strogatz.pdf")
def plot_barabasi_albert_graph():
"""
plot information in a pretty way
"""
plt.clf()
fig = plt.figure(1)
labels = [50, 100, 500, 1000, 2000, 5000]
x = [np.log(50), np.log(100), np.log(500),
np.log(1000), np.log(2000), np.log(5000)]
sparse_recov = [.35, .38, .58, .69, .79, .86]
max_likel = [.35, .38, .56, .68, .78, .85]
lasso = [.25, .3, .55, .67, .76, .83]
greedy = [.1, .13, .33, .47, .6, .75]
fig, ax = plt.subplots()
plt.axis((np.log(45), np.log(5500), 0, 1))
plt.xlabel("Number of Cascades")
plt.ylabel("F1 score")
plt.grid(color="lightgrey")
ax.plot(x, greedy, 'ko-', color="lightseagreen", label='Greedy')
ax.plot(x, lasso, 'ko-', color="orange", label="Lasso")
ax.plot(x, max_likel, 'ko-', color="cornflowerblue", label="MLE")
ax.plot(x, sparse_recov, 'ko-', color="k", label="Our Method")
plt.legend(loc="lower right")
ax.set_xticks(x)
ax.set_xticklabels(tuple(labels))
plt.savefig("../paper/figures/"+"barabasi_albert.pdf")
def plot_kronecker_l2norm():
plt.clf()
fig = plt.figure(1)
x = [50, 100, 500, 1000, 2000]
sparse_recov = [62, 60, 36, 28, 21]
max_likel = [139, 101, 42, 31, 25]
lasso = [50, 48, 33, 29, 23]
fig, ax = plt.subplots()
plt.subplots_adjust(bottom=.2, top=.85)
plt.xticks(ha="right", rotation=45)
plt.axis((50, 2000, 0, 145))
plt.xlabel("Number of Cascades")
plt.ylabel("l2-norm")
plt.grid(color="lightgrey")
ax.plot(x, lasso, 'ko-', color="orange", label="Lasso")
ax.plot(x, max_likel, 'ko-', color="cornflowerblue", label="MLE")
ax.plot(x, sparse_recov, 'ko-', color="k", label="Our Method")
plt.legend(loc="upper right")
ax.set_xticks(x)
ax.set_xticklabels(tuple(x))
plt.savefig("../paper/figures/"+"kronecker_l2_norm.pdf")
def plot_kronecker_l2norm_nonsparse():
plt.clf()
fig = plt.figure(1)
x = [50, 100, 500, 1000, 2000]
sparse_recov = [56, 55, 28, 21, 15]
max_likel = [125, 80, 35, 25, 20]
lasso = [47, 47, 27, 22, 17]
fig, ax = plt.subplots()
plt.subplots_adjust(bottom=.2, top=.85)
plt.xticks(ha="right", rotation=45)
plt.axis((50, 2000, 0, 145))
plt.xlabel("Number of Cascades")
plt.ylabel("l2-norm")
plt.grid(color="lightgrey")
ax.plot(x, lasso, 'ko-', color="orange", label="Lasso")
ax.plot(x, max_likel, 'ko-', color="cornflowerblue", label="MLE")
ax.plot(x, sparse_recov, 'ko-', color="k", label="Our Method")
plt.legend(loc="upper right")
ax.set_xticks(x)
ax.set_xticklabels(tuple(x))
plt.savefig("../paper/figures/"+"kronecker_l2_norm_nonsparse.pdf")
def plot_ROC_curve(figure_name):
"""
plot information in a pretty way
"""
plt.clf()
fig = plt.figure(2)
fig, ax = plt.subplots()
recall_sparse_200 = [.03, .16, .37, .4, .49]
precision_sparse_200 = [.9, .76, .61, .6, .63]
recall_lasso_200 = [.02, .11, .25, .43, .5, .54]
precision_lasso_200 = [.77, .77, .66, .56, .55, .51]
recall_sparse_50 = [.07, .13, .16, .58]
precision_sparse_50 = [.56, .53, .49, .37]
recall_lasso_50 = [.03, .18, .27, .82]
precision_lasso_50 = [.6, .47, .44, .24]
plt.xlabel("Recall")
plt.ylabel("Precision")
plt.grid(color="lightgrey")
ax.plot(recall_lasso_200, precision_lasso_200, 'ko-',
color="lightseagreen", label="Lasso-200 cascades")
ax.plot(recall_sparse_200, precision_sparse_200, 'ko-',
color="k", label="Our Method-200 cascades")
ax.plot(recall_lasso_50, precision_lasso_50, 'ko-',
color="orange", label="Lasso-50 cascades")
ax.plot(recall_sparse_50, precision_sparse_50, 'ko-',
color="cornflowerblue", label="Our Method-50 cascades")
plt.legend(loc="upper right", fontsize=14)
plt.savefig("../paper/figures/"+"ROC_curve.pdf")
def plot_kronecker_l2norm_nonsparse():
plt.clf()
fig = plt.figure(1)
x = [.01, .05, .1, .15, .2]
greedy = [.43, .29, .18, .1, .08]
sparse_recov = [.7, .58, .48, .39, .31]
max_likel = [.69, .56, .45, .37, .3]
lasso = [.66, .55, .46, .38, .3]
fig, ax = plt.subplots()
plt.subplots_adjust(bottom=.2, top=.85)
plt.xticks(ha="right", rotation=45)
plt.axis((0, .21, .05, .8))
plt.xlabel("Number of Cascades", fontsize=20)
plt.ylabel("l2-norm", fontsize=20)
plt.grid(color="lightgrey")
ax.plot(x, greedy, 'ko-', color="lightseagreen", label='Greedy')
ax.plot(x, lasso, 'ko-', color="orange", label="Lasso")
ax.plot(x, max_likel, 'ko-', color="cornflowerblue", label="MLE")
ax.plot(x, sparse_recov, 'ko-', color="k", label="Our Method")
plt.legend(loc="upper right", fontsize=18)
ax.set_xticks(x)
ax.set_yticklabels(tuple([.1, .2, .3, .4, .5, .6, .7, .8]), fontsize=20)
ax.set_xticklabels(tuple(x), fontsize=20)
plt.savefig("../paper/figures/"+"watts_strogatz_p_init.pdf")
if __name__=="__main__":
if 1:
compute_graph("../datasets/watts_strogatz_300_30_point3.txt",
n_cascades=300, lbda=.013382, min_proba=.2, max_proba=.7,
passed_function=
#convex_optimization.sparse_recovery)
algorithms.greedy_prediction, p_init=.2)
#convex_optimization.sparse_recovery, p_init=.15)
if 0:
compute_graph("../datasets/powerlaw_200_30_point3.txt",
n_cascades=200, lbda=.01, min_proba=.2, max_proba=.7,
passed_function=
#convex_optimization.sparse_recovery)
#algorithms.greedy_prediction)
convex_optimization.type_lasso)
if 0:
compute_graph("../datasets/barabasi_albert_300_30.txt",
n_cascades=100, lbda=.002, min_proba=.2,
max_proba=.7, passed_function=
convex_optimization.sparse_recovery)
#algorithms.greedy_prediction)
#convex_optimization.type_lasso)
if 0:
compute_graph("../datasets/kronecker_graph_256_cross.txt",
n_cascades=50, lbda=0., min_proba=.2, max_proba=.7,
passed_function=
convex_optimization.sparse_recovery,
#convex_optimization.type_lasso,
sparse_edges=True)
|
