An Incremental Learning Method Based on Dynamic Ensemble RVM for Intrusion Detection

Due to the dynamic changes of network data over time, static intrusion detection systems cannot adapt well behavioral characteristics input data, resulting in reduced accuracy. In addition, continuous streams will bring huge challenges resource storage and computing costs. Therefore, we propose an method ensemble incremental learning (DEIL-RVM), realize a dynamically adjusted model. which new overall misclassification probability weight value (OMPW) based on set or chunk is designed as basis for updating model, it can be used prune replace poor base component We presented probabilistic decision function taking into account posterior each RVM model dividing sample category. The with high sparsity obtain good balance between accuracy, robustness consumption, sacrifice less time cost while achieving higher accuracy stability streams.