Fuzzy Detection System for Rumors Through Explainable Adaptive Learning

Nowadays, rumor spreading has gradually evolved into a kind of organized behaviors, accompanied with strong uncertainty and fuzziness. However, existing fuzzy detection techniques for rumors focused their attention on supervised scenarios that require expert samples labels training. Thus, they are not able to well handle the unsupervised where unavailable. To bridge such gap, this article proposed system through explainable adaptive learning. Specifically, its core is graph embedding-based generative adversarial network (Graph-GAN) model. First all, it constructs fine-grained feature spaces via graph-level encoding. Furthermore, introduces continuous training between generator discriminator decoding. The two-stage scheme only solves under scenarios, but also improves robustness Empirically, set experiments carried out based three real-world datasets. Compared seven benchmark methods in terms four metrics, results Graph-GAN reveal proper performance, which averagely exceeds baselines by 5–10%.