It's always April fools' day!: On the difficulty of social network misinformation classification via propagation features

Abstract

Given the huge impact that Online Social Networks (OSN) had in the way people get informed and form their opinion, they became an attractive playground for malicious entities that want to spread misinformation, and leverage their effect. In fact, misinformation easily spreads on OSN and is a huge threat for modern society, possibly influencing also the outcome of elections, or even putting people's life at risk (e.g., spreading "anti-vaccines" misinformation). Therefore, it is of paramount importance for our society to have some sort of "validation" on information spreading through OSN. The need for a wide-scale validation would greatly benefit from automatic tools. In this paper, we show that it is difficult to carry out an automatic classification of misinformation considering only structural properties of content propagation cascades. We focus on structural properties, because they would be inherently difficult to be manipulated, with the the aim of circumventing classification systems. To support our claim, we carry out an extensive evaluation on Facebook posts belonging to conspiracy theories (as representative of misinformation), and scientific news (representative of fact-checked content). Our findings show that conspiracy content actually reverberates in a way which is hard to distinguish from the one scientific content does: for the classification mechanisms we investigated, classification F1-score never exceeds 0.65 during content propagation stages, and is still less than 0.7 even after propagation is complete.

The Digital Transformation of News Media and the Rise of Disinformation and Fake News

“…Intuitively, this is related to the difficulty of maintaining a higher number of ties, and the consequent weakness of the links. Conti et al (2017) provide additional empirical evidence reinforcing the hypothesis of echo chambers.…”

Natural Language Processing for Scam Detection. Classic and Alternative Analysis Techniques

“…As to those referred papers on Facebook data-sets based on machine learning analysis, Tacchini et al [67] showed that Facebook post can be highly accurately classified based on the users who "liked" them, obtaining classification accuracy over 99%. Additionally, Conti et al [13] focused on Online Social Network (OSN) structural properties of the information cascade as they are inherently difficult to be manipulated obtaining, on a highly imbalanced dataset and using a total of 28 features over three distinct classifiers (Linear Discriminant(LD), Random Forest (RF), Multi-Layer Perceptron (MLP)) classification F-score never exceeded 0.7.…”

“…On the text processing stage, it is lemmatized based on the SpaCy implementation and its' available large English dictionary 13 . Stop words, punctuation and spaces removal processes are also applied.…”

Identificación de noticias falsas sobre ciencia y tecnología por estudiantes del grado de Primaria

“…La tarea de identificarlas noticias falsas de forma temprana y automática con sistemas informáticos es crucial para no extender su difusión, sin embargo, algunos trabajos ya demuestran la complejidad que esto supone (Conti, Lain, Lazzeretti, Lovisotto, & Quattrociocchi, 2017;Del Vicario, Quattrociocchi, Scala, & Zollo, 2018). Como todavía los sistemas de detección de noticias falsas no son lo suficientemente sofisticados dado el volumen de las mismas, así como el peligro de las mismas para la sociedad en general (solo justificaría cualquier estudio el hecho de combatir el impacto que tienen estas noticias en la toma de decisión de muchos pacientes con problemas graves de salud que se acogen y automedican con remedios milagrosos de falsas noticias en los medios), es fundamental una formación para la ciudadanía sobre la capacidad de detección y evaluación de noticias falsas en cuanto a la información en general, y científica y tecnológica en particular, que difunden.…”