Decomposing one-class support vector machine into an ensemble of one-data support vector machines
Researchers have developed a new strategy to accelerate the performance of one-class support vector machines (OCSVMs), a common algorithm for anomaly detection. The proposed method involves decomposing large datasets into individual samples and training separate OCSVM models for each. These individual models are then combined using ensemble learning to create a comprehensive OCSVM model for the entire dataset. This approach, implemented in Python, has demonstrated faster processing times compared to traditional OCSVMs while maintaining similar classification accuracy. AI
IMPACT This research offers a method to improve the scalability of OCSVMs, potentially enabling more efficient anomaly detection on larger datasets.