Deep patch learning algorithms with high interpretability for regression problems

doi:10.1002/int.22937

CORC > 自动化研究所 > 中国科学院自动化研究所 > 复杂系统管理与控制国家重点实验室 > 先进控制与自动化团队

	Deep patch learning algorithms with high interpretability for regression problems
	Huang, Yunhu 4,5; Chen, Dewang 2,3,4; Zhao, Wendi 3,4; Lv, Yisheng 2; Wang, Shiping 1,5
刊名	INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
	2022-06-14
页码	38
关键词	deep learning deep patch learning fuzzy system fuzzy C-means clustering interpretability maximum information coefficient (MIC) Pearson's correlation coefficients (PCC)
ISSN号	0884-8173
DOI	10.1002/int.22937
通讯作者	Chen, Dewang(dwchen@fjut.edu.cn)
英文摘要	Improving the performance of machine learning algorithms to overcome the curse of dimensionality while maintaining interpretability is still a challenging issue for researchers in artificial intelligence. Patch learning (PL), based on the improved adaptive network-based fuzzy inference system (ANFIS) and continuous local optimization for the input domain, is characterized by high accuracy. However, PL can only handle low-dimensional data set regression. Based on the parallel and serial ensembles, two deep patch learning algorithms with embedded adaptive fuzzy systems (DPLFSs) are proposed in this paper. First, using the maximum information coefficient (MIC) and Pearson's correlation coefficients for feature selection, the variables with the least relationship (linear or nonlinear) are excluded. Second, principal component analysis is used to reduce the complexity further of DPLFSs. Meanwhile, fuzzy C-means clustering is used to enhance the interpretability of DPLFSs. Then, an improved PL method is put forward for the training of each sub-fuzzy system in a fashion of bottom-up layer-by-layer, and finally, the structure optimization is performed to significantly improve the interpretability of DPLFSs. Experiments on several benchmark data sets show the advantages of a DPLFS: (1) it can handle medium-scale data sets; (2) it can overcome the curse of dimensionality faced by PL; (3) its precision and generalization are greatly improved; and (4) it can overcome the poor interpretability of deep learning networks. Compared with shallow and deep learning algorithms, DPLFSs have the advantages of interpretability, self-learning, and high precision. DPLFS1 is superior for medium-scale data; DPLFS2 is more efficient and effective for high-dimensional problems, has a faster convergence, and is more interpretable.
资助项目	National Natural Science Foundation of China[61976055] ; Special Fund for Education and Scientific Research of Fujian Provincial Department of Finance[GY-Z21001] ; State Key Laboratory for Management and Control of Complex Systems[20210116]
WOS关键词	FUZZY SYSTEM ; UNIVERSAL APPROXIMATION
WOS研究方向	Computer Science
语种	英语
出版者	WILEY
WOS记录号	WOS:000810338500001
资助机构	National Natural Science Foundation of China ; Special Fund for Education and Scientific Research of Fujian Provincial Department of Finance ; State Key Laboratory for Management and Control of Complex Systems
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/49608]
专题	自动化研究所_复杂系统管理与控制国家重点实验室_先进控制与自动化团队
通讯作者	Chen, Dewang
作者单位	1.Fujian Prov Key Lab Network Comp & Intelligent In, Fuzhou, Peoples R China 2.Chinese Acad Sci, State Key Lab Management & Control Complex Syst, Inst Automat, Beijing, Peoples R China 3.Fujian Univ Technol, Intelligent Transportat Syst Res Ctr, Fuzhou, Peoples R China 4.FuJian Univ Technol, Sch Transportat, Fuzhou 350118, Peoples R China 5.Fuzhou Univ, Coll Comp & Data Sci, Fuzhou, Peoples R China
推荐引用方式 GB/T 7714	Huang, Yunhu,Chen, Dewang,Zhao, Wendi,et al. Deep patch learning algorithms with high interpretability for regression problems[J]. INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS,2022:38.
APA	Huang, Yunhu,Chen, Dewang,Zhao, Wendi,Lv, Yisheng,&Wang, Shiping.(2022).Deep patch learning algorithms with high interpretability for regression problems.INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS,38.
MLA	Huang, Yunhu,et al."Deep patch learning algorithms with high interpretability for regression problems".INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS (2022):38.