A Martingale Approach to Detect Peak of News in Social Network

TL;DR

This paper introduces a novel martingale-based statistical method for real-time detection of news peak points in social networks, enhancing understanding of information diffusion dynamics.

Contribution

It presents the first application of a martingale approach to online detection of news peaks in social media data, offering a new analytical tool.

Findings

Effective real-time detection of news peaks demonstrated

Good performance on real social media datasets

Potential for improved social network analysis

Abstract

Nowadays, social medias such as Twitter, Memetracker and Blogs have become powerful tools to propagate information. They facilitate quick dissemination sequence of information such as news article, blog posts, user's interests and thoughts through large scale. Providing strong means to analyzing social networks structure and how information diffuse through them is essential. Many recent studies emphasize on modeling information diffusion and their patterns to gain some useful knowledge. In this paper, we propose a statistical approach to online detect peak points of news when spread over social networks, to the best of our knowledge has never investigated before. The proposed model use martingale approach to predict peak points when news reached the peak of its popularity. Experimental results on real datasets show good performance of our approach to online detect these peak points.