Validating viral marketing strategies in Twitter via agent-based social simulation. (15th May 2016)

• Record Type: Journal Article
• Authors: Serrano, Emilio; Iglesias, Carlos A.
• Is Part Of: Expert systems with applications. Volume 50(2016)
• Page Start: 140
• Page End: 150
• Publication Date: 2016-05-15
• Subjects: Agent-based Social Simulation -- Viral marketing -- Social network analysis -- Rumor spreading model -- Twitter -- Big data; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2015.12.021
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: We present a method to validate viral marketing strategies in Twitter. The method is based on an Agent-based Social Simulation research methodology. A rumor spread model and control strategies are also presented as a case study. The model and the strategies are evaluated against others in the literature. An open-source simulator and the Twitter data employed are also given. Abstract: A number of marketing phenomena are too complex for conventional analytical or empirical approaches. This makes marketing a costly process of trial and error: proposing, imagining, trying in the real world, and seeing results. Alternatively, Agent-based Social Simulation (ABSS) is becoming the most popular approach to model and study these phenomena. This research paradigm allows modeling a virtual market to: design, understand, and evaluate marketing hypotheses before taking them to the real world. However, there are shortcomings in the specialized literature such as the lack of methods, data, and implemented tools to deploy a realistic virtual market with ABSS. To advance the state of the art in this complex and interesting problem, this paper is a seven-fold contribution based on a (1) method to design and validate viral marketing strategies in Twitter by ABSS. The method is illustrated with the widely studied problem of rumor diffusion in social networks. After (2) an extensive review of the related works for this problem, (3) an innovative spread model is proposed which rests on the exploratory data analysis of two different rumor datasets in Twitter. Besides, (4) new strategies are proposed to control malicious gossips. (5) The experimental results validate the realism of this new propagation model with the datasets and (6) the strategies performance is evaluated over this model. (7) Finally, the article is complemented by a free and open-source simulator.