Meta-cognitive online sequential extreme learning machine for imbalanced and concept-drifting data classification.

Journal: Neural networks : the official journal of the International Neural Network Society
PMID:

Abstract

In this paper, a meta-cognitive online sequential extreme learning machine (MOS-ELM) is proposed for class imbalance and concept drift learning. In MOS-ELM, meta-cognition is used to self-regulate the learning by selecting suitable learning strategies for class imbalance and concept drift problems. MOS-ELM is the first sequential learning method to alleviate the imbalance problem for both binary class and multi-class data streams with concept drift. In MOS-ELM, a new adaptive window approach is proposed for concept drift learning. A single output update equation is also proposed which unifies various application specific OS-ELM methods. The performance of MOS-ELM is evaluated under different conditions and compared with methods each specific to some of the conditions. On most of the datasets in comparison, MOS-ELM outperforms the competing methods.

Authors

  • Bilal Mirza
    School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore. Electronic address: bilal2@e.ntu.edu.sg.
  • Zhiping Lin
    School of Electrical and Electronic Engineering, Nanyang Technological University, Nanyang Avenue, 639798, Singapore. Electronic address: ezplin@ntu.edu.sg.