A Randomized Neural Network for Data Streams

被引:0
作者
Pratama, Mahardhika [1 ]
Angelov, Plamen P. [2 ]
Lu, Jie [3 ]
Lughofer, Edwin [4 ]
Seera, Manjeevan [5 ]
Lim, C. P. [6 ]
机构
[1] La Trobe Univ, Sch Engn & Math Sci, Melbourne, Vic, Australia
[2] Univ Lancaster, Sch Comp & Commun, Lancaster, England
[3] Univ Technol Sydney, Ctr Quantum Computat & Intelligent Syst, Sydney, NSW, Australia
[4] Johannes Kepler Univ Linz, Dept Knowledge Based Math Syst, Linz, Austria
[5] Swinburne Univ Technol, Fac Engn Comp & Sci, Sarawak Campus, Kuching, Malaysia
[6] Deakin Univ, Inst Intelligent Syst Res & Innovat, Geelong, Vic, Australia
来源
2017 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN) | 2017年
关键词
Evolving Fuzzy Systems; Fuzzy Neural Networks; Type-2 Fuzzy Systems; Sequential Learning; EXTREME LEARNING-MACHINE; RANDOM WEIGHT NETWORKS; ALGORITHMS;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Randomized neural network (RNN) is a highly feasible solution in the era of big data because it offers a simple and fast working principle in processing dynamic and evolving data streams. This paper proposes a novel RNN, namely recurrent type-2 random vector functional link network (RT2McRVFLN), which provides a highly scalable solution for data streams in a strictly online and integrated framework. It is built upon the psychologically inspired concept of metacognitive learning, which covers three basic components of human learning: what-to-learn, how-to-learn, and when-to-learn. The what-to-learn selects important samples on the fly with the use of online active learning scenario, which renders our algorithm an online semi-supervised algorithm. The how-to-learn process combines an open structure of evolving concept and a randomized learning algorithm of random vector functional link network (RVFLN). The efficacy of the RT2McRVFLN has been numerically validated through two real-world case studies and comparisons with its counterparts, which arrive at a conclusive finding that our algorithm delivers a tradeoff between accuracy and simplicity.
引用
收藏
页码:3423 / 3430
页数:8
相关论文
共 34 条
  • [31] A survey of randomized algorithms for training neural networks
    Zhang, Le
    Suganthan, P. N.
    [J]. INFORMATION SCIENCES, 2016, 364 : 146 - 155
  • [32] A local learning algorithm for random weights networks
    Zhao, Jianwei
    Wang, Zhihui
    Cao, Feilong
    Wang, Dianhui
    [J]. KNOWLEDGE-BASED SYSTEMS, 2015, 74 : 159 - 166
  • [33] Multivariable dynamic modeling for molten iron quality using online sequential random vector functional-link networks with self-feedback connections
    Zhou, Ping
    Yuan, Meng
    Wang, Hong
    Wang, Zhuo
    Chai, Tian-You
    [J]. INFORMATION SCIENCES, 2015, 325 : 237 - 255
  • [34] Active Learning With Drifting Streaming Data
    Zliobaite, Indre
    Bifet, Albert
    Pfahringer, Bernhard
    Holmes, Geoffrey
    [J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2014, 25 (01) : 27 - 39