Weakly supervised anomaly detection (WSAD) has gained increasing attention due to its core idea of enhancing the performance of unsupervised anomaly detection by leveraging prior knowledge from a limited number of labeled anomalies. In this article, we introduce a novel WSAD framework that surpasses current state-of-the-art methods in terms of accuracy, exhibits greater robustness to data uncertainty, and is more efficient in utilizing limited labeled anomalies. Our method is built upon twin fuzzy networks (TFN) that learn robust fuzzy if–then rules from a pairwise training set. TFN can extract informative prototypes of training instances, exploiting the very few labeled anomalies efficiently. A two-stage sequential training scheme, comprising fuzzy C-means clustering and interpolation consistency regularization, ensures that the fuzzy rules form a solid foundation for anomaly detection while improving TFN's generalization ability. The training process of TFN relies on closed-form optimization rather than gradient-based methods, leading to significantly faster training speeds. Comprehensive experiments conducted on numerous real-world datasets confirm the advantages of the TFN framework over existing alternatives.

In real-world data stream mining, the composition of classes undergoes unpredictable changes, giving rise to the challenge of class evolution, encompassing class emergence, disappearance, and reoccurrence. However, most existing approaches require the storage of past data to adapt their model. While some studies have focused on online learning approaches, they are built on an underlying assumption that the number of instances in any single class is consistently less than the sum of other classes. This assumption becomes invalid when a class emerges with a dominant amount, e.g., news about a pandemic outbreak, harming the performance of existing methods. In this paper, we thoroughly investigate this scenario and propose a novel online ensemble of ensemble one-versus-all framework (EEOF) to handle class evolution adaptively. The novel ensemble of ensemble architecture boosts diversity in each one-versus-all classifier. A novel adaptive model adaptation method is also designed to balance the error feedback between the emerging class and the other classes. A confidence-triggered fallback mode is integrated to prevent performance drop due to a wrong decision regarding class disappearance. Experimental studies are conducted on both synthetic and real-world data streams to show that our method achieves higher accuracy in diverse class evolution scenarios compared with the state-of-the-art method, particularly when classes emerge with dominant amounts.