Robust OS-ELM with a novel selective ensemble based on particle swarm optimization
This work addresses robustness and stability issues in online learning algorithms, particularly for applications using UCI datasets, but it is incremental as it builds upon existing OS-ELM methods.
The paper tackled the problem of improving robustness and stability in online sequential extreme learning machines by proposing a robust OS-ELM with a novel selective ensemble method based on particle swarm optimization, resulting in significant improvements as shown in experiments on UCI datasets for regression and classification.
In this paper, a robust online sequential extreme learning machine (ROS-ELM) is proposed. It is based on the original OS-ELM with an adaptive selective ensemble framework. Two novel insights are proposed in this paper. First, a novel selective ensemble algorithm referred to as particle swarm optimization selective ensemble (PSOSEN) is proposed. Noting that PSOSEN is a general selective ensemble method which is applicable to any learning algorithms, including batch learning and online learning. Second, an adaptive selective ensemble framework for online learning is designed to balance the robustness and complexity of the algorithm. Experiments for both regression and classification problems with UCI data sets are carried out. Comparisons between OS-ELM, simple ensemble OS-ELM (EOS-ELM) and the proposed ROS-ELM empirically show that ROS-ELM significantly improves the robustness and stability.