Research Area: Blockchain Technology
Memes are the thoughts, behaviors, or styles spread among people in the same cultural atmosphere, which keeps changing in semantics and emotion during the interaction with different individuals. With the rapid development of Internet technology, various network topics have emerged endlessly, making memes a cultural gene to interact and change during the propagation process frequently. In recent decades, several methods were proposed to simulate the extraction and tracing mechanism of the meme. Many dedicated evolutionary algorithms using meme theory were crafted to solve domain-specific complex problems more effectively. However, there are also some obvious shortcomings in the current research on meme prediction and discovery. Firstly, there is no central node for the propagation of meme in social networks, and the current research has not taken the meme propagation environment into account. Secondly, the existing models for meme prediction primarily use the dynamics model of virus spreading, which still lacks the study of modeling methods for meme spreading characteristics. In this paper, we present a scheme on decentralized blockchain theory, which is capable of discovering and predicting the transmission of the meme. A multi-agent theory is introduced to interpret the potential rules in a different agent and simulate the meme tracing in a decentralized environment. By comparing with widely used methods in the meme prediction experiment, the results demonstrate that the multi-agent model has the best prediction effect under three types of extracted features. We implement a prototype of Meme-chain and conducted experiments. The experimental results demonstrate that Meme-chain achieves excellent results on meme discovery and meme information transaction process with low latency and high accuracy. Actual case studies of the four types of meme discovery revealed that our proposed Meme-chain can be applied to actual social media data for meme discovery, with significant commercial value and research implications.
Keywords:
Author(s) Name: Fan Yang, Yanan Qiao, Shan Wang, Cheng Huang, Xiao Wang
Journal name: Knowledge-Based Systems
Conferrence name:
Publisher name: Elsevier
DOI: 10.1016/j.knosys.2021.107368
Volume Information: Volume 229, 11 October 2021, 107368
