
软件学报 ISSN 1000-9825, CODEN RUXUEW E-mail: jos@iscas.ac.cn
Journal of Software,2019,30(12):3637−3650 [doi: 10.13328/j.cnki.jos.005590] http://www.jos.org.cn
©中国科学院软件研究所版权所有. Tel: +86-10-62562563
一种面向中小规模数据集的模糊分类方法
∗
周
塔
1,2
,
邓赵红
1
,
蒋亦樟
1
,
王士同
1
1
(江南大学 数字媒体学院,江苏 无锡 214122)
2
(江苏科技大学 电子信息学院,江苏 张家港 215600)
通讯作者: 周塔, E-mail: jkdzhout@just.edu.cn
摘 要: 虽然 Takagi-Sugeno-Kang(TSK)模糊分类器在一些重要场合已经取得了广泛应用,但如何提高其分类性
能和增强其可解释性,仍然是目前的研究热点.提出一种随机划分与组合特征且规则具有高可解释性的深度 TSK 模
糊分类器(RCC-DTSK-C),但和其他分类器构造不同的是:(1) RCC-DTSK-C 由很多基训练单元构成,这些基训练单
元可以被独立训练;(2) 每一个基训练单元的隐含层通过模糊规则的可解释性来表达,而这些模糊规则又是通过随
机划分、随机组合来进行特征选择的;(3) 基于栈式结构理论,源数据集作为相同的输入空间被映射到每一个独立
的基训练单元中,这样就有效地保证了源数据的所有特征在每一个独立的训练单元中都得以保留.实验结果表明,
RCC-DTSK-C 具有良好的分类性能和可解释性.
关键词: Takagi-Sugeno-Kang(TSK);随机模糊划分;特征组合;可解释性;深度学习;栈式结构
中图法分类号: TP18
中文引用格式: 周塔,邓赵红,蒋亦樟,王士同.一种面向中小规模数据集的模糊分类方法.软件学报,2019,30(12):3637−3650 .
http://www.jos.org.cn/1000-9825/5590.htm
英文引用格式: Zhou T, Deng ZH, Jiang YZ, Wang ST. Fuzzy classification method for small- and medium-scale datasets. Ruan
Jian Xue Bao/Journal of Software, 2019,30(12):3637−3650 (in Chin ese). h ttp://www.jos .org.cn/1000-9825/5590 .htm
Fuzzy Classi fication Me thod for Small- and Me dium- scale Datase ts
ZHOU Ta
1,2
, DENG Zhao-Hong
1
, JIANG Yi-Zhang
1
, WANG Shi-Tong
1
1
(School of Digital Media, Jiangnan University, Wuxi 214122, Chin a)
2
(School of Electronics and Information, Jiangsu University of Science and Technology, Zhangjiagang 215600, China)
Abstra ct : Although Takagi-Sugeno-Kang (TSK) is widely used in practically every profession, how to enhance its classification
accuracy and interpretability is still a research focus. In this study, a deep TSK fuzzy classifier is proposed. This classifier (i.e., RCC-
DTSK-C) can randomly select features and combine features and own triplely concise interpretability for fuzzy rules. There are several
other varieties of RCC-DTSK-C such as reasonable structure for rule representation, namely, (1) the proposed RCC-DTSK-C consists of
many base-training units and each base-training unit can be trained independently. According to the principle of stacked generalization,
the input of the next base-training unit consists of the training set and random result obtained from random projections about prediction
results of current base-training unit. (2) In RCC-DTSK-C, the hidden layer of each base-training unit is represented by triplely concise
interpretable fuzzy rules which are in the sense of randomly selected features. These features are selected by dividing into the not-fixed
several fuzzy partitions and randomly combining rules and keeping the same input space in every base-training unit. (3) The source data
set is mapped into each of the independent base-training units as the same input space, which effectively ensures that all the features of
the source data are preserved in each separate training unit. The extensive experimental results show RCC-DTSK-C can achieve the
enhanced classification performance and triplely concise interpr etability for fuzzy rules.
∗ 基金项目: 国家自然科学基金(61772239, 61702225, 61572236, 61711540041)
Foundation item: National Natural Science Foundation of China (6177223 9, 61702225, 61572236, 617 11540041)
收稿时间: 2017-0 9-17; 修改时间: 2018-04-16; 采用时间: 2018-04-23; jos 在线出版时间: 2019-01-21
CNKI 网络优先出版: 2019-01-22 13:49:03, http://kns.cnki.net/kcms/d etail/11.2560.TP.20190122.1348.011.html
评论