Quantitative detection of disinformation patterns using semantic vector analysis

This study examines the semantic coherence of disinformation headlines using publicly available machine learning tools. By applying sentiment, temporal, and security classifiers, the research confirms that syntactically distinct but thematically aligned headlines cluster in a shared semantic space. The findings support Claire Wardle’s concept of information environments and highlight the enduring importance of human judgment in combating information manipulation.