Towards an automatic system for fake news detection

Comments and information appearing on the internet and on different social media sway opinion concerning potential remedies for diagnosing and curing diseases. The article we present today, elaborates an algorithm to analyze the behavior of users of Twitter, the most important social network with respect to this issue, as well as a dynamic knowledge graph construction method based on information gathered from Twitter and other open data sources such as web pages.

The article has been prepared by Pablo Lara Navarra, member of the Management & e-Learning (MeL) research group, together with Hervé Falciani, Enrique A. Sánchez Pérez and Antonia Ferrer Sapena. This was published in the International Journal of Environmental Research and Public Health in 2020.

About the article

The article presents a methodology for constructing a graph database of tweets using the program Neo4j to analyze the dissemination of false information, particularly in the context of public health issues. The structured database allows for the visualization of tweet chains and provides a heuristic tool for analysts to detect false information. The study demonstrates a general analytical approach to studying the dissemination of socially sensitive information, focusing on rapid diffusion, unknown sources, and a large audience, such as in the case of non-standard drugs, environmental issues, and the anti-vaccination movement.

The proposed platform could serve as a basis for heuristic analysis of dynamic information like fake news and potentially support the development of automatic algorithms for detecting false information in healthcare.

The study concludes by suggesting that while the current methodology is not yet fully automated for detecting false information, it represents a significant step towards achieving this goal in the future.

Consult the study

You can read the full article by clicking here.