Dynamic Rankings for Seed Selection in Complex Networks: Balancing Costs and Coverage

Jankowski, Jarosław

doi:10.3390/e19040170

Cited by 12 publications

(8 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The presented results lay firm foundation for the earlier studies related to sequential seeding and address currently unanswered research questions [21][22][23]. Averaging the coverage obtained by the typical agent based simulations delivers improvement in about 90% of the cases [21], the formal proof relying on the coordinated execution and corroborated by the new empirical studies shows that sequential seeding delivers at least as large coverage as the widely used single stage seeding; under modest assumption this coverage is provably larger.…”

Section: Discussionsupporting

confidence: 73%

“…It takes the advantage of delayed seeding by interweaving activating seeds with diffusion. Here we firm the foundations of sequential seeding and address a number of research questions that have not been studied before [21][22][23]. First, we prove formally and confirm experimentally on real networks that the sequential seeding is at least as good as the single stage method and under modest conditions, it is provably better.…”

Section: Introductionmentioning

confidence: 63%

See 1 more Smart Citation

Probing Limits of Information Spread with Sequential Seeding

Jankowski

Szymański

Kazienko

et al. 2018

Sci Rep

Self Cite

View full text Add to dashboard Cite

We consider here information spread which propagates with certain probability from nodes just activated to their not yet activated neighbors. Diffusion cascades can be triggered by activation of even a small set of nodes. Such activation is commonly performed in a single stage. A novel approach based on sequential seeding is analyzed here resulting in three fundamental contributions. First, we propose a coordinated execution of randomized choices to enable precise comparison of different algorithms in general. We apply it here when the newly activated nodes at each stage of spreading attempt to activate their neighbors. Then, we present a formal proof that sequential seeding delivers at least as good spread coverage as the single stage seeding does. Moreover, we also show that, under modest assumptions, sequential seeding performs provably better than the single stage seeding using the same number of seeds and node ranking. Finally, we present experimental results comparing single stage and sequential approaches on directed and undirected graphs to the well-known greedy approach to provide the objective measure of the sequential seeding benefits. Surprisingly, applying sequential seeding to a simple degree-based selection leads to higher coverage than achieved by the computationally expensive greedy approach currently considered to be the best heuristic.

show abstract

Section: Discussionsupporting

confidence: 73%

Section: Introductionmentioning

confidence: 63%

Probing Limits of Information Spread with Sequential Seeding

Jankowski

Szymański

Kazienko

et al. 2018

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…These types of solutions are based more on better use of processes of natural diffusion and use sequential seeding [27], avoid nodes from within the same communities with intra connections that are close by using target communities [28], use dynamic rankings with sequential seeding [29] and use mechanisms for voting that have lower weights once activated nodes have been detected [30]. Apart from basic centrality measures, the central nodes in networks can be detected using a k-shell based approach [31].…”

Section: Methodsmentioning

confidence: 99%

Multi-criteria decision support for planning and evaluation of performance of viral marketing campaigns in social networks

2018

Self Cite

View full text Add to dashboard Cite

The current marketing landscape, apart from conventional approaches, consists of campaigns designed especially for launching information diffusion processes within online networks. Associated research is focused on information propagation models, campaign initialization strategies and factors affecting campaign dynamics. In terms of algorithms and performance evaluation, the final coverage represented by the fraction of activated nodes within a target network is usually used. It is not necessarily consistent with the real marketing campaigns using various characteristics and parameters related to coverage, costs, behavioral patterns and time factors for overall evaluation. This paper presents assumptions for a decision support system for multi-criteria campaign planning and evaluation with inputs from agent-based simulations. The results, which are delivered from a simulation model based on synthetic networks in a form of decision scenarios, are verified within a real network. Last, but not least, the study proposes a multi-objective campaign evaluation framework with several campaign evaluation metrics integrated. The results showed that the recommendations generated with the use of synthetic networks applied to real networks delivered results according to the decision makers’ expectation in terms of the used evaluation criteria. Apart from practical applications, the proposed multi-objective approach creates new evaluation possibilities for theoretical studies focused on information spreading processes within complex networks.

show abstract

“…Sequential seeding was further extended towards the recalculation of rankings with effective degree used in each seeding stage [36]. In another study, the ability of buffering not used seeds was analysed with seeds collected in buffer during no seeding stages.…”

Section: Sequential Seeding For Temporal Networkmentioning

confidence: 99%

Effective Influence Spreading in Temporal Networks With Sequential Seeding

2020

Self Cite

View full text Add to dashboard Cite

The spread of influence in networks is a topic of great importance in many application areas. For instance, one would like to maximise the coverage, limiting the budget for marketing campaign initialisation and use the potential of social influence. To tackle this and similar challenges, more than a decade ago, researchers started to investigate the influence maximisation problem. The challenge is to find the best set of initially activated seed nodes in order to maximise the influence spread in networks. In typical approach we will activate all seeds in single stage, at the beginning of the process, while in this work we introduce and evaluate a new approach for seeds activation in temporal networks based on sequential seeding. Instead of activating all nodes at the same time, this method distributes the activations of seeds, leading to higher ranges of influence spread. The results of experiments performed using real and randomised networks demonstrate that the proposed method outperforms single stage seeding in 71% of cases by nearly 6% on average. Knowing that temporal networks are an adequate choice for modelling dynamic processes, the results of this work can be interpreted as encouraging to apply temporal sequential seeding for real world cases, especially knowing that more sophisticated seed selection strategies can be implemented by using the seed activation strategy introduced in this work.

show abstract

Dynamic Rankings for Seed Selection in Complex Networks: Balancing Costs and Coverage

Cited by 12 publications

References 42 publications

Probing Limits of Information Spread with Sequential Seeding

Probing Limits of Information Spread with Sequential Seeding

Multi-criteria decision support for planning and evaluation of performance of viral marketing campaigns in social networks

Effective Influence Spreading in Temporal Networks With Sequential Seeding

Contact Info

Product

Resources

About