It's Always April Fools' Day! On the Difficulty of Social Network Misinformation Classification via Propagation Features

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