D2H-AD: A Hybrid Model Utilizing Hyperdimensional Computing for Advanced Anomaly Detection

In three linesD2H-AD is an anomaly detection framework based on Hyperdimensional Computing (HDC). It combines density-aware encoding and distance-based similarity, outperforming five baselines (HDAD, ODHD, One-Class SVM, Isolation Forest, Autoencoders) across five datasets. Hyperdimensional encoding alone achieves +5.4% ROC-AUC improvement. Lightweight, interpretable, computationally efficient: suited for TinyML and edge AI.

Summary generated by Claude — human-verified