RIMR: Reverse Influence Maximization Rank

The ubiquitous usage of social media and other internet platforms has led to the widespread proliferation of naturally occurring social networks. This has caused a revolution in marketing and information dissemination due to the rise of influencers, and their innate ability to broadcast information to large audiences. Influence maximization, a key problem in identifying optimal influential individuals, can be used to identify influencers in a social network, or can be superimposed on other settings which exhibit relational information. Popular methods, such as greedy algorithms, face scalability issues, performing poorly on larger networks, and incurring a large memory overhead. To address these challenges, this paper introduces a novel methodology, RIMR, which leverages statistical modeling and parallel programming to circumvent the need for storing large collections of data, achieving scalability and efficiency. Experimental results demonstrate the effectiveness and scalability of the proposed approach compared to state-of-the-art methods, showcasing significant speedups. This paper provides insights into the development of the proposed statistical process, the heteroge-neous algorithm implementation, and a prototype demonstrating its efficacy.