source code for FCM-RDpA (paper|arxiv|code|blog)
FCM-RDpA (Fuzzy C-Means Clustering, Regularization, DropRule, and Powerball AdaBelief) enhances the MBGD-RDA (Mini-Batch Gradient Descent with Regularization, DropRule, and AdaBound; paper|code|blog) in the following three aspects for TSK fuzzy regression model construction.
run demoAA.m to reproduce the results on the Concrete-CS dataset of Fig.3/4/8 in the paper.
run demoPS.m to reproduce the results on the Concrete-CS dataset of Fig.5 in the paper.
run demoInit.m to reproduce the results on the Concrete-CS dataset of Fig.6 in the paper.
run demoGD.m to reproduce the results on the Concrete-CS dataset of Fig.7 in the paper.
@Article{Shi2021,
author = {Zhenhua Shi and Dongrui Wu and Chenfeng Guo and Changming Zhao and Yuqi Cui and Fei-Yue Wang},
journal = {Information Sciences},
title = {{FCM-RDpA}: {TSK} Fuzzy Regression Model Construction Using Fuzzy C-Means Clustering, Regularization, {D}rop{R}ule, and {P}owerball {A}da{B}elief},
year = {2021},
pages = {490-504},
volume = {574},
}
@Article{Wu2020,
author = {Dongrui Wu and Ye Yuan and Jian Huang and Yihua Tan},
journal = {IEEE Trans. on Fuzzy Systems},
title = {Optimize {TSK} Fuzzy Systems for Regression Problems: Mini-batch Gradient Descent With Regularization, {D}rop{R}ule, and {A}da{B}ound ({MBGD-RDA})},
year = {2020},
number = {5},
pages = {1003-1015},
volume = {28},
}