Professional Documents
Culture Documents
Computational Data and Social Networks 7Th International Conference Csonet 2018 Shanghai China December 18 20 2018 Proceedings Xuemin Chen
Computational Data and Social Networks 7Th International Conference Csonet 2018 Shanghai China December 18 20 2018 Proceedings Xuemin Chen
https://textbookfull.com/product/big-data-analytics-6th-
international-conference-bda-2018-warangal-india-
december-18-21-2018-proceedings-anirban-mondal/
https://textbookfull.com/product/advanced-data-mining-and-
applications-14th-international-conference-adma-2018-nanjing-
china-november-16-18-2018-proceedings-guojun-gan/
https://textbookfull.com/product/theory-and-practice-of-natural-
computing-7th-international-conference-tpnc-2018-dublin-ireland-
december-12-14-2018-proceedings-david-fagan/
https://textbookfull.com/product/computational-science-
iccs-2018-18th-international-conference-wuxi-china-
june-11-13-2018-proceedings-part-ii-yong-shi/
https://textbookfull.com/product/computational-science-
iccs-2018-18th-international-conference-wuxi-china-
june-11-13-2018-proceedings-part-iii-yong-shi/
https://textbookfull.com/product/proceedings-of-international-
conference-on-computational-intelligence-and-data-engineering-
proceedings-of-iccide-2018-nabendu-chaki/
Xuemin Chen · Arunabha Sen
Wei Wayne Li · My T. Thai (Eds.)
LNCS 11280
Computational Data
and Social Networks
7th International Conference, CSoNet 2018
Shanghai, China, December 18–20, 2018
Proceedings
123
Lecture Notes in Computer Science 11280
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology Madras, Chennai, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7407
Xuemin Chen Arunabha Sen
•
Computational Data
and Social Networks
7th International Conference, CSoNet 2018
Shanghai, China, December 18–20, 2018
Proceedings
123
Editors
Xuemin Chen Wei Wayne Li
Texas Southern University Texas Southern University
Houston, TX, USA Houston, TX, USA
Arunabha Sen My T. Thai
Ira A. Fulton School of Engineering University of Florida
Tempe, AZ, USA Gainesville, AL, USA
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
This book constitutes the refereed proceedings of the 7th International Conference on
Computational Data and Social Networks Social Networks, CSoNet 2018, held in
Shanghai, China, during December 18–20, 2018. This conference provides a premier
interdisciplinary forum to bring together researchers and practitioners from all fields of
complex networks, i.e., social networks, communication networks, and senor networks,
working on various fundamental and emerging issues such as network computing,
modeling, social network analysis, and data mining for presentation of original research
results, as well as the exchange and dissemination of innovative, practical development
experiences.
The topics cover the fundamental background, theoretical technology development,
and real-world applications associated with complex and data network analysis, min-
imizing influence of rumors on social networks, blockchain Markov modeling, fraud
detection, data mining, Internet of Things (IoT), Internet of Vehicles (IoV), and others.
Selected papers will be published in special issues of Journal of Combinatorial
Optimization, IEEE Transactions on Network Science and Engineering, and Compu-
tational Social Networks.
The intended audience of this book mainly consists of researchers, research stu-
dents, and practitioners in complex networks and data mining, focused more on social
networks and wireless sensor networks. The book is also of interest to researchers and
industrial practitioners in emerging areas such as artificial intelligence, big data,
blockchain, cloud computing, fraud detection, intelligent transportation system, spam
detection, and rumor blocker.
We would like to express our appreciation to all contributors and the conference
committee members. A special appreciation goes to Shanghai Polytechnic University
(SPU) for its support of this conference. Furthermore, without the unlimited efforts
from the local chair, Dr. Chuang Ma, and his team members, publicity co-chairs,
Drs. Xiang Li, Mugeng Peng, and Wenan Tan, financial chair, Dr. Thang Dinh, and
website chair, David Smith, our conference would not have been so success in its
seventh anniversary year. We would also like to recognize Dr. Jian Wang’s contri-
bution in editing the final proceedings document. In addition, we sincerely appreciate
the consistent support and great effort in the publication process of these proceedings
by Anna Kramer and Alfred Hofmann from Springer.
General Chairs
Wei Wayne Li Texas Southern University, USA
My T. Thai University of Florida, USA
Publicity Co-chairs
Xiang Li Santa Clara University, USA
Mugen Peng Beijing University of Posts and Telecommunications, China
Wenan Tan Shanghai Polytechnic University, China
Financial Chair
Thang Dinh Virginia Commonwealth University, USA
Web Chair
J. David Smith University of Florida, USA
Contents
The Design of Real Time Recording System for Electric Spark Source
Parameters Based on the Internet of Things . . . . . . . . . . . . . . . . . . . . . . . . 310
Mingzhang Luo, Hao Zhang, Qiang Liu, Yang Xu, Chong Luo,
and Zhongyuan Wang
Short Papers
1 Introduction
With the advance of the internet and computer technology, some significant
social networks have been widely integrated into our daily life, such as Face-
book, Twitter and Google+. Social networks can usually be represented as com-
plex networks of nodes and edges, where nodes denote the individuals (people,
organizations, or other social entities) and edges denote the social relationships
This work is partly supported by National Natural Science Foundation of China under
grant 11671400 and National Science Foundation under grant 1747818.
c Springer Nature Switzerland AG 2018
X. Chen et al. (Eds.): CSoNet 2018, LNCS 11280, pp. 1–12, 2018.
https://doi.org/10.1007/978-3-030-04648-4_1
2 R. Yan et al.
The rest of this paper is organized as follows. We first begin by recalling some
existing related work of rumor blocking in Sect. 2. Then we introduce information
diffusion model in Sect. 3. And we show the problem description and properties
in Sect. 4. Algorithm is presented in Sect. 5. We analyze and discuss the results
of the experiments in Sect. 6. Finally, we draw our conclusions in Sect. 7.
2 Related Work
Domingos et al. [4] first study the influence between users for marketing in social
networks. Kempe et al. [7] model viral marketing as a discrete optimization prob-
lem, which is named Influence Maximization (IM). They propose a greedy algo-
rithm with (1 − 1/e)-approximation ratio since the function is submodular under
Independent Cascade (IC) or Linear Threshold (LT) model. Based on Kempe’s
contributions, there have been substantial efforts in modeling the propagation
of information in recent years such as [15–17].
In [6], Fan et al. study the problem that identifies a minimal subset of individuals
as initial protectors (the nodes are used to limit the bad influence of rumors.)
to minimize the number of people infected in neighbor communities at the end
of both diffusion processes. Authors propose algorithms under Opportunistic
One-Activate-One and Deterministic One-Activate-Many models and show the
theoretical analysis in detail. In [14], Wang et al. address the problem of mini-
mizing the influence of rumor. In their paper, they assume a rumor emerges in
the social network and part of users have already adopted it, their goal is to
minimize the size of ultimately contaminated users by discovering and blocking
k uninfected users. A greedy method for efficiently finding a good approximate
solution to this problem is proposed. Unfortunately, they have no theoretical
analysis. In social networks, how to identify the influential spreaders is crucial
for rumors. Ma et al. in [10] propose a gravity centrality index to identify the
influential spreaders in complex networks and compare with some well-known
centralities, such as degree, betweenness, closeness, and so forth.
In [9], Kimura et al. propose a method (by blocking a limited number of links) for
efficiently finding a good approximate solution to rumor blocking. In [8], Khalil
et al. propose a edge deletion problem and optimize the topology of a networked
4 R. Yan et al.
In [2], Budak et al. propose the problem that identifying a subset of individuals
that need to be convinced to adopt the competing (or good) campaign so as to
minimize the number of people that adopt the bad campaign at the end of both
propagation processes. And they show this problem is NP-hard and provide a
greedy algorithm. In [12], Tong et al. study the rumor blocking problem that
asks for k seed users to trigger the spread of a positive cascade such that the
number of the users who are not influenced by rumor can be maximized. They
present a randomized approximation algorithm which is provably superior to the
state-of-the art methods with respect to running time.
⎧
⎪ 1, if v ∈ S
⎪
⎪
⎨ 0, if N in (v) = ∅
P rM (v, S) = (1)
⎪
⎪ (1 − P rM (u, S)puv ), otherwise.
⎪
⎩ 1−
u∈N in (v)
Where N in (v) is the set of in-neighbors of v and P rM (u, S)puv represents the
probability u successfully activates v under the diffusion model M (Here, M is
IC model). As we can clearly see the activation probability of a node v depends
on the its in-neighbors u. Then we give the problem description as follow
Definition 1. Minimizing Influence of Rumor (MIR). Given a directed social
network G = (V, E, p), a rumor source set S, a positive integer budget k, and
the IC model M, MIR aims to find a blocker set B with k nodes such that
B ∗ = arg min P rM (v, S). (2)
B⊆V \S,|B|=k
v∈V \{S∪B}
5 Algorithm
In this section, we propose a two-stages method GCSSB which includes generat-
ing candidate set and selecting blockers stages. We introduce them in Sects. 5.1
and 5.2 respectively.
Fig. 2. An example with propagation probability p = 0.5. Figure 2(a) shows the adja-
cent matrix where Aij = 0.5 means there is a directed edge from i to j. Figure 2(b)
shows how to calculate σ.
Figure 2(b) shows σ = I + AI + ...Ar I = (3.938, 3.531, 2.5, 1.5, 1.5, 1, 3.656)T
where r = 5. And we sort σ in descending order. Then we obtain permutation
Π = (3.938, 3.656, 3.531, 2.5, 1.5, 1.5, 1)T and choose the top α ∗ k nodes as the
candidate set of blockers C where α is a threshold parameter. Consistent with
the example mentioned earlier, we choose node 1 instead of node 3 because
σ1 = 3.938 > σ3 = 2.5 when k = 1.
6 Experiment
In this section, we evaluate proposed algorithm on synthetic and real-life net-
works. First, we describe the data sets and experiment setup. Second, we analyze
and discuss experimental results from different perspectives. Finally, we compare
with other heuristic approaches.
8 R. Yan et al.
We generate a random network and collect three real-life social networks with
various scale from Stanford Large Network Dataset Collection (SNAP)1 and
the Koblenz Network Collection (KONECT)2 respectively. Table 1 provides the
details of these data sets. In table, ’CC’ represents clustering coefficient and ’MD’
represents the maximum degree.
We make the following setup for rumor spreading process: Given a directed
social network G = (V, E, p), 1% of nodes are selected randomly and uniformly
from V as rumor source set S. In our all experiments, we adopt Independent
Cascade (IC) model as information diffusion model. In particular, we assign p
in two ways since the data sets lack propagation probability p. One assigns a
uniform probability p = 0.5 for each edge on the networks. Another assigns a
trivalency model p = T RI for each edge, i.e., we uniformly select a value from
{0.1,0.01,0.01} at random that corresponds to high, medium and low propagation
probabilities. Notice that all networks are simple networks3 .
3
Self-loops and multiple edges are not allowed.
10 R. Yan et al.
6.3 Results
Fig. 3. Total activation probability Vs. Parameter α under IC model: rumor source
|S| = 1%|V | on each network, p = 0.5 or p = T RI, and k = 50.
Fig. 4. Total activation probability Vs. Parameter k under IC model: rumor source
|S| = 1%|V | on each network, p = 0.5 or p = T RI, and α = 6.
Minimizing Influence of Rumors by Blockers on Social Networks 11
Fig. 5. Compare with other methods: rumor source |S| = 1%|V |, p = 0.5 on TL
network or p = T RI on G+ network, and α = 6.
Parameter k Study: We study the relationship between the size of the blocker
set and the total activation probability. The experimental results are shown in
Fig. 4. In the subfigures, the horizontal and vertical axes represent the parame-
ter k and the total activation probability, respectively. Through experiments, we
observe that the total activation probability decreases as k increases. In partic-
ular, it is drastically reduced when k > 300 on each network.
Comparing with Other Methods: We compare our GCSSB with other meth-
ods (OD, BC and PR). The experimental results are shown in Fig. 5. The hori-
zontal and vertical axes represent the parameter k and the total activation proba-
bility, respectively. In both subfigures, The total activation probability decreases
as k increases. We observe that the proposed method is the best since the total
activation probability is the smallest. Moreover, in comparison methods, the
PR’s performance is the best but the OD is the worst.
7 Conclusions
References
1. Brandes, U.: On variants of shortest-path betweenness centrality and their generic
computation. Soc. Netw. 30(2), 136–145 (2008)
2. Budak, C., Agrawal, D., El Abbadi, A.: Limiting the spread of misinformation in
social networks. In: Proceedings of the 20th International Conference on World
Wide Web, pp. 665–674. ACM (2011)
3. Doerr, B., Fouz, M., Friedrich, T.: Why rumors spread so quickly in social networks.
Commun. ACM 55(6), 70–75 (2012)
4. Domingos, P., Richardson, M.: Mining the network value of customers. In: Pro-
ceedings of the Seventh ACM SIGKDD International Conference on Knowledge
Discovery and Data Mining, pp. 57–66. ACM (2001)
5. ERDdS, P., R&WI, A.: On random graphs i. Publ. Math. Debrecen 6, 290–297
(1959)
6. Fan, L., Lu, Z., Wu, W., Thuraisingham, B., Ma, H., Bi, Y.: Least cost rumor block-
ing in social networks. In: 2013 IEEE 33rd International Conference on Distributed
Computing Systems (ICDCS), pp. 540–549. IEEE (2013)
7. Kempe, D., Kleinberg, J., Tardos, É.: Maximizing the spread of influence through
a social network. In: Proceedings of the Ninth ACM SIGKDD International Con-
ference on Knowledge Discovery and Data Mining, pp. 137–146. ACM (2003)
8. Khalil, E.B., Dilkina, B., Song, L.: Scalable diffusion-aware optimization of network
topology. In: Proceedings of the 20th ACM SIGKDD International Conference on
Knowledge Discovery and Data Mining, pp. 1226–1235. ACM (2014)
9. Kimura, M., Saito, K., Motoda, H.: Minimizing the spread of contamination by
blocking links in a network. In: AAAI, vol. 8, pp. 1175–1180 (2008)
10. Ma, L.L., Ma, C., Zhang, H.F., Wang, B.H.: Identifying influential spreaders in
complex networks based on gravity formula. Phys. A Stat. Mech. Appl. 451, 205–
212 (2016)
11. Page, L., Brin, S., Motwani, R., Winograd, T., et al.: The PageRank citation
ranking: bringing order to the web (1998)
12. Tong, G., et al.: An efficient randomized algorithm for rumor blocking in online
social networks. IEEE Trans. Netw. Sci. Eng. (2017)
13. Tong, H., Prakash, B.A., Eliassi-Rad, T., Faloutsos, M., Faloutsos, C.: Gelling,
and melting, large graphs by edge manipulation. In: Proceedings of the 21st ACM
International Conference on Information and Knowledge Management, pp. 245–
254. ACM (2012)
14. Wang, S., Zhao, X., Chen, Y., Li, Z., Zhang, K., Xia, J.: Negative influence minimiz-
ing by blocking nodes in social networks. In: AAAI (Late-Breaking Developments),
pp. 134–136 (2013)
15. Wen, S., Haghighi, M.S., Chen, C., Xiang, Y., Zhou, W., Jia, W.: A sword with
two edges: propagation studies on both positive and negative information in online
social networks. IEEE Trans. Comput. 64(3), 640–653 (2015)
16. Yan, R., Zhu, Y., Li, D., Ye, Z.: Minimum cost seed set for threshold influence
problem under competitive models. World Wide Web 1–20 (2018)
17. Zhu, Y., Li, D., Zhang, Z.: Minimum cost seed set for competitive social influence.
In: IEEE INFOCOM 2016-The 35th Annual IEEE International Conference on
Computer Communications, pp. 1–9. IEEE (2016)
Budgeted Competitive Influence
Maximization on Online Social Networks
1 Introduction
Online social network (OSN) has recently been a very effective mean in diffus-
ing information, propagating opinions or ideas. Many companies have leveraged
word-of-mouth effect in OSNs to promote their products. The key problem of
viral marketing is Influence Maximization (IM), which aims to select a set of
k users (called seed set) in a social networks with maximum influence spread.
Kempe et al. [8] first formulated IM problem in two diffusion models, named Lin-
ear Threshold (LT) and Independent Cascade (IC), which simulated the propaga-
tion of influence through social networks. This problem has been widely studied
due to its important role in viral marketing [2,10,14,16,17].
Table 1. Notations
Notations Descriptions
n, m the number of nodes and the number of edges
N− (v), N+ (v) the sets of incoming, and outgoing neighbor nodes of v
I(·), U(·) The expected number of A-active nodes, and its upper bound, respectively
S ∗ , SU
∗
Optimal solution for BCIM, maximum I(·), and U(·)
OPT, OPTu I(S ∗ ), U(SU
∗
)
2 2 1
Υ (, δ) (1 + )(2 + 3 ) ln δ 2
CovR (S) the number of URR sets Rj be covered by S
1
Û(S) Û(S) = |R|
CovR (S) an estimation of U over URR sets R
kmax max{k : ∃A ⊆ V, c(A) ≤ L}
16 C. V. Pham et al.
– At step t = 0, A0 = SA , B0 = SB .
– At step t ≥ 1, each node v ∈/ At−1 ∩ Bt−1 becomes A-active if
wA (u, v) ≥ θA (v) and wB (u, v) < θB (v) (1)
u∈N− (v)∩At−1 u∈N− (v)∩Bt−1
We follow the method in [7] to construct a live-edge model and prove this model
is equivalent to TCLT model.
Live-Edge Construction. From original graph G = (V, E) and weights wA
and wB , respectively, we construct sample graph (or realization) g from G as
follows: For each v ∈ V , we randomly select one in-edge (u, v) with
probability
wA (u, v), and do not select any in-edge with probability 1 − u∈V wA (u, v).
The selected edge is called A-edge. On the other hand, we also randomly select
one in-edge (u, v) (called B-edge)with probability wB (u, v), and do not select
any in-edge with probability 1 − u∈V wB (u, v). Let gA , gB be the sub-graph
including only A-edges and B-edges, respectively. Finally, we return g as union
of gA and gB .
A-active Nodes and B-active Nodes Distribution on Live-Edge Model.
On a live-edge g, we denote At and Bt are set of A-active and B-active nodes on
g at step t. The distribution of A-active and B-active nodes in g was be happen
in discrete steps t as follows: