A Novel Hybrid Approach for Influence Maximization in Online Social Networks Based on Node Neighborhoods

Nandi, Gora Chand; Sharma, Uzzal; Das, Anjan

doi:10.1007/978-981-10-4765-7_54

Cited by 3 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chen et al [6] exploited a degree discount heuristic algorithm, which nearly matched the performance of the greedy algorithms for the IC model, and improved upon the pure degree heuristic in the other cascade models. Nandi et al [7] proposed a new method called DegGreedy to maximize the influence spread based on node neighborhoods, which could provide higher influence spread and good efficiency in terms of scalability. Deng et al [8] proposed two centrality-based edge activation probability algorithms under the IC model, which named NewDiscount and GreedyCIC, with considering edge probability.…”

Section: Related Workmentioning

confidence: 99%

Influence Maximization with Consideration of PageRank Centrality and Propagation Probability

Qi¹

2021

Proceedings of the 33rd International Conference on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

Influence Maximization (IM) problem has been attracted considerable interest and attention in last decades. However, the centrality algorithm-based methods were with low time complexity but made the acceptability of diffusion vaguely. The main purpose of our work is to select the influential nodes according to the available budget to maximize the impact coverage. Based on the traditional independent cascade model, this paper mainly solves the IM problem, designs two effective PRTH algorithms based on PageRank and propagation probability threshold, and combines PageRank of PRTH processed graph with degree discount algorithm to get an algorithm named PRDD. Experiments on four datasets show that the two algorithms have better performance than the existing algorithms in the aspect of influence diffusion.

show abstract

Section: Related Workmentioning

confidence: 99%

Influence Maximization with Consideration of PageRank Centrality and Propagation Probability

Qi¹

2021

Proceedings of the 33rd International Conference on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

show abstract

“…The authors provided a simple greedy algorithm that evaluates all the nodes in the network and combination of this method with a Monte Carlo simulation is, however computationally expensive. While a series of follow-up works have suggested additional developments such as a combination with evolutionary algorithms to improve computational times, these approaches are still limited in terms of performance, and strictly applied to conventional social networks [14,15]. Another category of approach has applied heuristic algorithms with the advantage of better running time, but still limited to some approximations.…”

Section: Related Workmentioning

confidence: 99%