HDSVM: A High Efficiency Distributed SVM Framework over Data Stream

The application of Support Vector Machine (SVM) over data stream is growing with the increasing real-time processing requirements in classification field, like anomaly detection and real-time image processing. However, the dynamic live data with high volume and fast arrival rate in data streams make it challenging to apply SVM in data stream processing. Existing SVM implementations are mostly designed for batch processing and hardly satisfy the...