A Review of Online Sequential Extreme Learning Machines
Chang Hui Deng, Xiao Jun Wang, Jun Gu, Wei Wang · Journal of Physics Conference Series · 2019
Abstract One of the challenges for machine leaning on big data is the effective and efficient leaning of large-scale and on-going explosion data which is always with the concept drift problem. To meet the challenge, learning algorithms/techniques performed well on large-scale data and also with the evolvable property are desired. The OS-ELM family has strong potential as viable alternative techniques for the computation of large-scale and on-going explosion data in more fields of applications/tasks. This work reviews the most important and latest works in OS-ELM family. The review consists of two topics, one related to the improved version of OS-ELM which aims at overcoming the disadvantages of OS-ELM, and the other related to the extended version the goals of which is to add some specialties to OS-ELM. It is expected that the review will support a certain research in the future.