You are on page 1of 6

e-ISSN (O): 2348-4470

Scientific Journal of Impact Factor (SJIF): 4.72


p-ISSN (P): 2348-6406

International Journal of Advance Engineering and Research


Development
Volume 4, Issue 2, February -2017

A Query Approach For Influence Maximization On Specific User In Social


Network.
Anuj Mishra1, Rahul Kumar Deo2, Ashutosh3
1
Department of Computer Engineering, Indira College of Engineering, Pune
2
Department of Computer Engineering, Indira College of Engineering, Pune
3
Department of Computer Engineering, Indira College of Engineering, Pune

Abstract Influence maximization is introduced to maximize the profit of viral promoting in social networks. The
weakness of influence maximization is that it does not distinguish specific users from others, even if some items may be
solely helpful for the specific users. For such items, it is a much better strategy to specialize in maximizing the influence
on the specific users. In this paper, we formulate Associate in Nursing influence maximization problem as question
process to distinguish specific users from others. We propose an expectation model for the price of the target operate and
a quick greedy-based approximation method mistreatment the expectation model. For the expectation model, we
investigate a relationship of paths between users. For the greedy method, we work out an economical progressive change
of the marginal gain to our objective operate. Our experimental results show that (a) our improved greedy algorithm
achieves better running time comparing with the development of with matching influence unfold, (b) our degree discount
heuristics achieve a lot of higher influence unfold than classic degree and centrality-based heuristics, and when tuned for
a specific influence cascade model, it achieves almost matching influence thread with the greedy algorithmic program,
and more significantly.

Keyword: Graph algorithms, influence maximization, independent cascade model, social networks.

I. INTRODUCTION

As of late, the measure of unfold of knowledge is unrelentingly enlarged in on-line social organizations, for example,
Facebook and Twitter. To utilize online social organizations as a promoting stage, there are masses of examination on the
most effective thanks to utilize the proliferation of impact for infective agent advertising. One of the exploration issues is
influence maximization (IMAX), which plans to discover k seed purchasers to amplify the unfold of impact among
purchasers in social organizations. It is ended up being a NP-hard issue by Kempe et al. Since they proposed associate
greedy calculation for the issue, numerous analysts have planned completely different heuristic routines.
Viral showcasing is one of the key utilizations of impact boost. In viral advertising, a thing that associate publicist desires
to advance is subtle into informal communities "by overhearing people's conversations" correspondence. From the point
of read of advertising, impact augmentation gives however to get the foremost extreme have the benefit of each one of
the purchasers in a casual organization through infective agent showcasing. In any case, impact amplification is not
generally the most effective technique for infective agent showcasing, on the grounds that there can be some things that
are useful to simply specific purchasers. These particular purchasers will be one or two people with a typical enthusiasm
for a given factor, some or all individuals in a cluster, or some or all clients in a category. There is no restriction for being
particular purchasers. For instance, consider associate publicist that is approached to advance a restorative item for
women through infective agent showcasing. For the corrective item, the particular purchasers are feminine purchasers
why ought to seemingly utilize it and male purchasers WHO want to shop for it as a gift for feminine purchasers. For this
situation, the advertiser will not ought to be distressed concerning alternate purchasers in lightweight of the actual fact
that the restorative item isn't useful to them. Rather, it is a superior method to focus on augmenting the amount of
wedged specific purchasers, yet impact amplification has the disadvantage that it cannot acknowledge them from
alternate purchasers. The main method for taking care of such focuses with impact boost is creating a unvaried diagram
with the objectives and capital punishment impact growth on the chart. On the other hand, the aftereffect of this
methodology got to be off base, on the grounds that there will be some purchasers WHO don't seem to be targets
however rather can unambiguously impact the objectives.

II. LITERATURE SURVEY

1. Maximizing the spread of influence through a social network


AUTHORS: D. Kempe, J. Kleinberg, and E. Tardos
Models for the processes by which concepts and influence propagate through a social network have been studied during a
range of domains, including the diffusion of medical and technological innovations, the sudden and widespread adoption
of numerous methods in game-theoretic settings, and the effects of "word of mouth" within the promotion of latest
products. Recently, motivated by the style of microorganism promoting methods, Domingos and Richardson posed a
@IJAERD-2017, All rights Reserved 239
International Journal of Advance Engineering and Research Development (IJAERD)
Volume 4, Issue 2, February -2017, e-ISSN: 2348 - 4470, print-ISSN: 2348-6406

basic recursive drawback for such social network processes: if we have a tendency to will try and win over a set of
people to adopt a brand new product or innovation, and the goal is to trigger an oversized cascade of further adoptions,
which set of people ought to we have a tendency to target? we have a tendency to think about this drawback in many of
the foremost wide studied models in social network analysis. The optimization drawback of choosing the foremost
powerful nodes is NP-hard here, and we offer the primary demonstrable approximation guarantees for economical
algorithms. Using associate analysis framework primarily based on sub standard functions, we show that a natural greedy
strategy obtains a answer that's most likely inside sixty three of optimum for many categories of models; our framework
suggests a general approach for reasoning concerning the performance guarantees of algorithms for these styles of
influence issues in social networks. We additionally offer procedure experiments on massive collaboration networks,
showing that in addition to their provable guarantees, our approximation algorithms significantly out-perform node-
selection heuristics primarily based on the well-studied notions of degree spatial relation and distance spatial relation
from the sphere of social networks.

2. Labeled influence maximization in social networks for target marketing


AUTHORS: F.-H. Li, C.-T. Li, and M.-K. Shan
The influence maximization problem is to notice a collection of seed nodes that maximize the unfold of influence in a
very social network. The seed nodes are used for the microorganism selling to gain the utmost profits through the
effective viva-voce. However, in more real-world cases, marketers usually target bound merchandise at specific teams of
customers. While original influence maximization downside considers no product data and target customers, in this
paper, we focus on the target selling. We propose the labelled influence maximization downside, which aims to notice a
collection of seed nodes which might trigger the utmost unfold of influence on the target customers in a very labelled
social network. We propose 3 algorithms to solve such labelled influence maximization downside. We 1st develop the
algorithms based mostly on the greedy strategies of original influence maximization by considering the target customers.
Moreover, we develop a novel algorithmic rule, Maximum Coverage, whose central idea is to offline cipher the try wise
proximities of nodes in the labelled social network and on-line notice the set of seed nodes. This allows the marketers to
arrange and assess methods on-line for publicized merchandise. The experimental results on IMDb labeled social
network show our strategies will accomplish promising performances on each effectiveness and potency.

3. Profit maximization over social networks


AUTHORS: W. Lu and L. Lakshmanan
Influence maximization is the problem of finding a collection of prestigious users in an exceedingly social network such
the expected unfold of influence below an explicit propagation model is maximized. Much of the previous work has
neglected the necessary distinction between social influence and actual product adoption. However, as recognized in the
management science literature, an individual WHO gets influenced by social acquaintances might not essentially adopt a
product (or technology), due, e.g., to monetary issues. In this work, we distinguish between influence and adoption by
expressly modeling the states of being influenced and of adopting a product. We extend the classical Linear Threshold
(LT) model to incorporate costs and valuations, and factor them into users' decision-making method of adopting a
product. We show that the expected profit operate below our projected model maintains submodularity below sure
conditions, but no longer exhibits monotonicity, unlike the expected influence unfold operate. To maximize the expected
profit under our extended LT model, we use associate degree unbudgeted greedy framework to propose 3 profit
maximization algorithms. The results of our detailed experimental study on 3 real-world datasets demonstrate that of the
3 algorithms, PAGE, which assigns costs dynamically primarily based on the profit potential of every candidate seed, has
the best performance both within the expected profit achieved and in period.

4. Mining the network value of customers.


AUTHORS: P. Domingos and M. Richardson,
One of the foremost applications of information mining is in helping firms verify that potential customers to promote to.
If the expected profit from a customer is bigger than the price of selling to her, the marketing action for that client is
dead. So far, work in this area has thought-about solely the intrinsic price of the client (i.e, the expected profit from sales
to her). We propose to model conjointly the customer's network value: the expected profit from sales to alternative
customers she could influence to shop for, the customers those may influence, and so on recursively. Instead of viewing a
market as a group of independent entities, we read it as a social network and model it as a Andre Markoff random field.
We show the benefits of this approach employing a social network well-mined from a cooperative filtering info.
Marketing that exploits the network price of customers---also acknowledged as microorganism marketing---can be very
effective, but is still a necromancy. Our work can be viewed as a step towards providing a additional solid foundation for
it, taking advantage of the availability of huge relevant databases.

5. Community-based Greedy Algorithm for Mining Top-K Influential Nodes in Mobile Social Networks
AUTHORS: Yu Wang, Gao ong, Guojie Song, Kunqing Xie
It is a fundamental issue to search out a set of authoritative people during a mobile social network specified targeting
them at the start can maximize the unfold of the influence. The problem of finding the foremost authoritative nodes is

@IJAERD-2017, All rights Reserved 240


International Journal of Advance Engineering and Research Development (IJAERD)
Volume 4, Issue 2, February -2017, e-ISSN: 2348 - 4470, print-ISSN: 2348-6406

sadly NP-hard. It has been shown that a Greedy algorithm with demonstrable approximation guarantees will offer
sensible approximation; but, it is computationally expensive, if not prohibitive, to run the greedy algorithm on a giant
mobile network. In this paper we propose a replacement algorithmic rule known as Community based mostly Greedy
algorithmic rule for mining top-K authoritative nodes. The proposed algorithmic rule encompasses 2 components: 1)
associate degree algorithmic rule for police work communities in a social network by taking under consideration info
diffusion; and 2) a dynamic programming algorithmic rule for choosing communities to search out authoritative nodes.
We conjointly offer demonstrable approximation guarantees for our algorithmic rule.

III. PROPOSED SYSTEM

We propose a replacement economical expectation model for the influence unfold of a seed set supported freelance
most influence ways in which (IMIP) among users. we tend to jointly show that the new objective perform of the
new expectation model is sub normal.
Based on the new expectation model, we tend to gift a technique to efficiently methodology associate IMAX
question. the tactic consists of distinctive native regions containing nodes that influence the target nodes of a
question and approximating best seeds from the native regions as a result of the results of the question. distinctive
such native regions helps to cut back the interval, once the number of targets in associate IMAX question is no
compared to the quantity of all nodes.

We by experimentation experiment by experimentation through associate experiment demonstrate that our


distinctive native influencing regions technique is extraordinarily powerful and thus the planned technique may be a
minimum of an order of magnitude faster than the comparison ways in which in most cases with high accuracy.
Distinctive native influencing regions makes the essential greedy algorithmic rule relating to cardinal times faster at
intervals the experiments.

Figure: Planned System Design

@IJAERD-2017, All rights Reserved 241


International Journal of Advance Engineering and Research Development (IJAERD)
Volume 4, Issue 2, February -2017, e-ISSN: 2348 - 4470, print-ISSN: 2348-6406

IV. Mathematical Model


Let S is the Whole System Consist of
S= {I, P, O}
I = Input.
I = {U, Q, D}
U = User
U = {u1,u2.un}
Q = Query Entered by user
Q = {q1, q2, q3qn}
D = Dataset.
P = Process:
P = {IMIP, CELF, PMIA, IRIE, CD Model}
IMIP: independent maximum influence paths
CELF: Cost-Effective Lazy Forward
PMIA: prefix excluding maximum influence arborescence
Step1: User enters the Query.
Step2: In comparison process following methods will be performed.
Step3: Independent maximum influence paths (IMIP):
We propose a new efficient expectation model for the influence spread of a seed set based on independent maximum
influence paths (IMIP) among users.
Step4: CELF++: is an improved greedy algorithm exploiting Sub-modularity.
Step5: PMIA is a greedy-based algorithm based on maximum influence paths between nodes. In PMIA, parameter u is
used to prune out maximum influence
paths having low influence .
Step6: IRIE: is one of recent algorithms for influence maximization.
Step7: CD Model: CD is the greedy method using the CD model. The CD model is a probabilistic model based on users
historical action logs. We use this method only for the experiment related to the actual influence spread.
Output: Finally the particular result will be shown to user as per his query.

V. RESULT AND DISCUSSIONS

1. Home window

@IJAERD-2017, All rights Reserved 242


International Journal of Advance Engineering and Research Development (IJAERD)
Volume 4, Issue 2, February -2017, e-ISSN: 2348 - 4470, print-ISSN: 2348-6406

2. User Registration Window

3. User Login Window:

4. Result Window:

@IJAERD-2017, All rights Reserved 243


International Journal of Advance Engineering and Research Development (IJAERD)
Volume 4, Issue 2, February -2017, e-ISSN: 2348 - 4470, print-ISSN: 2348-6406

VI. CONCLUSION

In this paper, we detail IMAX question making ready to expand the impact on specific purchasers in informal
organizations. Since IMAX inquiry handling is NP-hard and ascertaining its target capacity is #P-hard, we concentrate on
the foremost skilful technique to inexact ideal seeds effectively. To estimate the target's estimation capacity, we propose
the IMIP model in read of freedom between ways that. To prepare an IMAX question proficiently, removing possibility
for ideal seeds is projected and the fast ravenous based mostly estimate utilizing the IMIP model.

ACKNOWLEDGMENT

We might want to thank the analysts and also distributers for making their assets accessible. We additionally appreciative
to commentator for their significant recommendations furthermore thank the school powers for giving the obliged base
and backing.

REFRENCES

[1] Jong-Ryul Lee and Chin-Wan Chung A Query Approach for Influence Maximization on Specific Users in Social
Networks

[2] F.-H. Li, C.-T. Li, and M.-K. Shan, Labeled influence maximization in social networks for target marketing, in
Proc. IEEE 3rd Int. Conf. Privacy, Secur., Risk Trust, Int. Conf. Social Comput., 2011, pp. 560563.

[3] W. Lu and L. Lakshmanan, Profit maximization over social networks, in Proc. IEEE 12th Int. Conf. Data Mining,
2012, pp. 479488.

[4] P. Domingos and M. Richardson, Mining the network value of customers, in Proc. 7th ACM SIGKDD Int. Conf.
Knowl. Discovery Data Mining, 2001, pp. 5766.

[5] J. Leskovec, A. Krause, C. Guestrin, C. Faloutsos, J. VanBriesen, and N. Glance, Cost-effective outbreak detection
in networks, in Proc. 13th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, 2007, pp. 420429.

[6] A. Goyal, W. Lu, and L. V. Lakshmanan, CELF++: Optimizing the greedy algorithm for influence maximization in
social networks, in Proc. 20th Int. Conf. Companion World Wide Web, 2011, pp. 4748.

[7] Y. Wang, G. Cong, G. Song, and K. Xie, Community-based greedy algorithm for mining top-k influential nodes in
mobile social networks, in Proc. 16th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, 2010, pp. 1039
1048.

[8] W. Chen, Y. Wang, and S. Yang, Efficient influence maximization in social networks, in Proc. 15th ACM
SIGKDD Int. Conf. Knowl. Discovery Data Mining, 2009, pp. 199208.

[9] W. Chen, C. Wang, and Y. Wang, Scalable influence maximization for prevalent viral marketing in large-scale
social networks, in Proc. 16th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, 2010, pp. 10291038.

[10] Q. Jiang, G. Song, C. Gao, Y. Wang, W. Si, and K. Xie, Simulated annealing based influence maximization in
social networks, in Proc. 25th AAAI Conf. Artif. Intell., 2011, pp. 127132.

@IJAERD-2017, All rights Reserved 244