Fuzzy-rough set based unsupervised dynamic feature selection algorithm

doi:10.11772/j.issn.1001-9081.2022101543

Journal of Computer Applications ›› 2023, Vol. 43 ›› Issue (10): 3121-3128.DOI: 10.11772/j.issn.1001-9081.2022101543

Special Issue: 数据科学与技术

• Data science and technology • Previous Articles Next Articles

Fuzzy-rough set based unsupervised dynamic feature selection algorithm

Lei MA¹, Chuan LUO¹(), Tianrui LI², Hongmei CHEN²

^1.College of Computer Science，Sichuan University，Chengdu Sichuan 610065，China
^2.School of Computing and Artificial Intelligence，Southwest Jiaotong University，Chengdu Sichuan 611756，China

Received:2022-10-17 Revised:2022-11-28 Accepted:2022-11-30 Online:2023-10-07 Published:2023-10-10
Contact: Chuan LUO
About author:MA Lei， born in 1997， M. S. candidate. His research interests include feature selection.
LI Tianrui， born in 1969， Ph. D.， professor. His research interests include data mining， granular computing.
CHEN Hongmei， born in 1971， Ph. D.， professor. Her research interests include data mining， granular computing.
Supported by:
National Natural Science Foundation of China(62076171);Natural Science Foundation of Sichuan Province(2022NSFSC0898)

基于模糊粗糙集的无监督动态特征选择算法

马磊¹, 罗川¹(), 李天瑞², 陈红梅²

^1.四川大学计算机学院，成都 610065
^2.西南交通大学计算机与人工智能学院，成都 611756

通讯作者: 罗川
作者简介:马磊（1997—），男（回族），四川成都人，硕士研究生，主要研究方向：特征选择
李天瑞（1969—），男，福建莆田人，教授，博士，CCF会员，主要研究方向：数据挖掘、粒计算
陈红梅（1971—），女，四川成都人，教授，博士，CCF会员，主要研究方向：数据挖掘、粒计算。
基金资助:
国家自然科学基金资助项目(62076171);四川省自然科学基金资助项目(2022NSFSC0898)

Abstract

Abstract:

Dynamic feature selection algorithms can improve the time efficiency of processing dynamic data. Aiming at the problem that there are few unsupervised dynamic feature selection algorithms based on fuzzy-rough sets， an Unsupervised Dynamic Fuzzy-Rough set based Feature Selection （UDFRFS） algorithm was proposed under the condition of features arriving in batches. First， by defining a pseudo triangular norm and new similarity relationship， the process of updating fuzzy relation value was performed on the basis of existing data to reduce unnecessary calculation. Then， by utilizing the existing feature selection results， dependencies were adopted to judge if the original feature part would be recalculated to reduce the redundant process of feature selection， and the feature selection was further speeded up. Experimental results show that compared to the static dependency-based unsupervised fuzzy-rough set feature selection algorithm， UDFRFS can achieve the time efficiency improvement of more than 90 percentage points with good classification accuracy and clustering performance.

Key words: feature selection, fuzzy-rough set, dynamic data, unsupervised feature selection, dependency

摘要：

动态特征选择算法能够大幅提升处理动态数据的效率，然而目前基于模糊粗糙集的无监督的动态特征选择算法较少。针对上述问题，提出一种特征分批次到达情况下的基于模糊粗糙集的无监督动态特征选择（UDFRFS）算法。首先，通过定义伪三角范数和新的相似关系在已有数据的基础上进行模糊关系值的更新过程，从而减少不必要的运算过程；其次，通过利用已有的特征选择结果，在新的特征到达后，使用依赖度判断原始特征部分是否需要重新计算，以减少冗余的特征选择过程，从而进一步提高特征选择的速度。实验结果表明，UDFRFS相较于静态的基于依赖度的无监督模糊粗糙集特征选择算法，在时间效率方面能够提升90个百分点以上，同时保持较好的分类精度和聚类表现。

关键词: 特征选择, 模糊粗糙集, 动态数据, 无监督特征选择, 依赖度

CLC Number:

TP301.6

Lei MA, Chuan LUO, Tianrui LI, Hongmei CHEN. Fuzzy-rough set based unsupervised dynamic feature selection algorithm[J]. Journal of Computer Applications, 2023, 43(10): 3121-3128.

马磊, 罗川, 李天瑞, 陈红梅. 基于模糊粗糙集的无监督动态特征选择算法[J]. 《计算机应用》唯一官方网站, 2023, 43(10): 3121-3128.

Figures/Tables 8

References 22

1	PAWLAK Z. Rough sets［J］. International Journal of Computing and Information Sciences， 1982， 11（5）： 341-356. 10.1007/bf01001956
2	DUBOIS D， PRADE H. Rough fuzzy sets and fuzzy rough sets［J］. International Journal of General Systems， 1990， 17（2/3）： 191-209. 10.1080/03081079008935107
3	ZADEH L A. Fuzzy sets［J］. Information and Control， 1965， 8（3）： 338-353. 10.1016/s0019-9958(65)90241-x
4	AN S， LIU J， WANG C， et al. A relative uncertainty measure for fuzzy rough feature selection［J］. International Journal of Approximate Reasoning， 2021， 139： 130-142. 10.1016/j.ijar.2021.09.014
5	CHEN J， MI J， LIN Y. A graph approach for fuzzy-rough feature selection［J］. Fuzzy Sets and Systems， 2020， 391： 96-116. 10.1016/j.fss.2019.07.014
6	YANG X， CHEN H， LI T， et al. A noise-aware fuzzy rough set approach for feature selection［J］. Knowledge-Based Systems， 2022， 250： No.109092. 10.1016/j.knosys.2022.109092
7	章夏杰，朱敬华，陈杨. Spark下的分布式粗糙集属性约简算法［J］. 计算机应用， 2020， 40（2）： 518-523.
	ZHANG X J， ZHU J H， CHEN Y. Distributed rough set attribute reduction algorithm under Spark［J］. Journal of Computer Applications， 2020， 40（2）： 518-523.
8	危前进，魏继鹏，古天龙，等. 粗糙集多目标并行属性约简算法［J］. 软件学报， 2022， 33（7）： 2599-2617.
	WEI Q J， WEI J P， GU T L， et al. Multi-objective parallel attribute reduction algorithm in rough set［J］. Journal of Software， 2022， 33（7）： 2599-2617.
9	YANG Y， CHEN D， WANG H， et al. Fuzzy rough set based incremental attribute reduction from dynamic data with sample arriving［J］. Fuzzy Sets and Systems， 2017， 312： 66-86. 10.1016/j.fss.2016.08.001
10	WAN J， CHEN H， LI T， et al. Dynamic interaction feature selection based on fuzzy rough set［J］. Information Sciences， 2021， 581： 891-911. 10.1016/j.ins.2021.10.026
11	NI P， ZHAO S， WANG X， et al. Incremental feature selection based on fuzzy rough sets［J］. Information Sciences， 2020， 536： 185-204. 10.1016/j.ins.2020.04.038
12	程玉胜，陈飞，王一宾. 基于粗糙集的数据流多标记分布特征选择［J］. 计算机应用， 2018， 38（11）： 3105-3111， 3118. 10.11772/j.issn.1001-9081.2018041275
	CHENG Y S， CHEN F， WANG Y B. Feature selection for multi-label distribution learning with streaming data based on rough set［J］. Journal of Computer Applications， 2018， 38（11）： 3105-3111， 3118. 10.11772/j.issn.1001-9081.2018041275
13	曾艺祥，林耀进，范凯钧，等. 基于层次类别邻域粗糙集的在线流特征选择算法［J］. 南京大学学报（自然科学版）， 2022， 58（3）： 506-518.
	ZENG Y X， LIN Y J， FAN K J， et al. Online streaming feature selection method based on hierarchical class neighborhood rough set［J］. Journal of Nanjing University （Natural Sciences）， 2022， 58（3）： 506-518.
14	SOLORIO-FERNÁNDEZ S， CARRASCO-OCHOA J A， MARTÍNEZ-TRINIDAD J F. A review of unsupervised feature selection methods［J］. Artificial Intelligence Review， 2020， 53（2）： 907-948. 10.1007/s10462-019-09682-y
15	VELAYUTHAM C， THANGAVEL K. Unsupervised quick reduct algorithm using rough set theory［J］. Journal of Electronic Science and Technology， 2011， 9（3）： 193-201.
16	VELAYUTHAM C， THANGAVEL K. A novel entropy based unsupervised feature selection algorithm using rough set theory［C］// Proceeding of the 2012 International Conference on Advances in Engineering， Science and Management. Piscataway： IEEE， 2012： 156-161. 10.1504/ijcbdd.2012.045949
17	PARTHALÁIN N MAC， JENSEN R. Unsupervised fuzzy-rough set-based dimensionality reduction［J］. Information Sciences， 2013， 229： 106-121. 10.1016/j.ins.2012.12.001
18	YUAN Z， CHEN H， LI T， et al. Unsupervised attribute reduction for mixed data based on fuzzy rough sets［J］. Information Sciences， 2021， 572： 67-87. 10.1016/j.ins.2021.04.083
19	GANIVADA A， RAY S S， PAL S K. Fuzzy rough sets， and a granular neural network for unsupervised feature selection［J］. Neural Networks， 2013， 48： 91-108. 10.1016/j.neunet.2013.07.008
20	BEG I， ASHRAF S. Similarity measures for fuzzy sets［J］. Applied and Computational Mathematics， 2009， 8（2）： 192-202.
21	PAP E. Pseudo-analysis in engineering decision making［M］// RUDAS I J， FODOR J， KACPRZYK J. Towards Intelligent Engineering and Information Technology， SCI 243. Berlin： Springer， 2009： 3-16. 10.1007/978-3-642-03737-5_1
22	WANG L， PENG L， ZHANG G. Attribute reduction for incomplete and unsupervised data based on fuzzy rough sets［C］// Proceeding of the 16th International Conference on Intelligent Systems and Knowledge Engineering. Piscataway： IEEE， 2021： 674-679. 10.1109/iske54062.2021.9755349

数据集	样本数	特征数	聚类数	来源
Olive Oil	30	570	4	TSC
SCADI	70	205	7	UCI
SAP	131	21	3	UCI
Divorce	170	54	2	UCI
Sonar	208	60	2	UCI
Ionosphere	351	33	2	UCI
Wdbc	569	30	2	UCI
German	1 000	20	2	UCI
Waveform	5 000	21	3	UCI

数据集	样本数	特征数	聚类数	来源
Olive Oil	30	570	4	TSC
SCADI	70	205	7	UCI
SAP	131	21	3	UCI
Divorce	170	54	2	UCI
Sonar	208	60	2	UCI
Ionosphere	351	33	2	UCI
Wdbc	569	30	2	UCI
German	1 000	20	2	UCI
Waveform	5 000	21	3	UCI

数据集	UDFRFS	UFRFS		FRUAR		SUIFR
数据集	运行时间/s	运行时间/s	TRR/%	运行时间/s	TRR/%	运行时间/s	TRR/%
Olive Oil	0.001 1	0.496 6	0.21	7.739 2	0.01	50.603 8	0.00
SCADI	0.005 5	0.484 2	1.13	134.344 8	0.00	21.391 2	0.03
SAP	0.001 8	0.035 7	5.07	1.457 5	0.12	1.009 1	0.18
Divorce	0.004 1	0.186 5	2.19	17.875 9	0.02	10.665 2	0.04
Sonar	0.005 7	0.294 4	1.92	11.802 5	0.05	48.826 6	0.01
Ionosphere	0.022 1	0.300 4	7.37	32.265 6	0.07	111.249 9	0.02
Wdbc	0.039 4	0.704 8	5.60	25.644 7	0.15	98.890 0	0.04
German	0.115 0	1.749 1	6.57	98.004 1	0.12	58.844 5	0.20
Waveform	4.086 7	58.758 3	6.96	—	—	—	—

数据集	UDFRFS	UFRFS		FRUAR		SUIFR
数据集	运行时间/s	运行时间/s	TRR/%	运行时间/s	TRR/%	运行时间/s	TRR/%
Olive Oil	0.001 1	0.496 6	0.21	7.739 2	0.01	50.603 8	0.00
SCADI	0.005 5	0.484 2	1.13	134.344 8	0.00	21.391 2	0.03
SAP	0.001 8	0.035 7	5.07	1.457 5	0.12	1.009 1	0.18
Divorce	0.004 1	0.186 5	2.19	17.875 9	0.02	10.665 2	0.04
Sonar	0.005 7	0.294 4	1.92	11.802 5	0.05	48.826 6	0.01
Ionosphere	0.022 1	0.300 4	7.37	32.265 6	0.07	111.249 9	0.02
Wdbc	0.039 4	0.704 8	5.60	25.644 7	0.15	98.890 0	0.04
German	0.115 0	1.749 1	6.57	98.004 1	0.12	58.844 5	0.20
Waveform	4.086 7	58.758 3	6.96	—	—	—	—

数据集	UDFRFS	UFRFS		FRUAR		SUIFR
数据集	运行时间/s	运行时间/s	TRR/%	运行时间/s	TRR/%	运行时间/s	TRR/%
OliveOil	0.003 1	0.852 2	0.37	12.747 0	0.02	84.319 0	0.00
SCADI	0.011 2	0.855 2	1.31	230.457 1	0.00	37.209 8	0.03
SAP	0.004 7	0.049 7	9.41	2.698 4	0.17	1.754 5	0.27
Divorce	0.009 2	0.292 9	3.14	28.547 7	0.03	16.738 9	0.05
Sonar	0.012 7	0.498 5	2.54	19.326 7	0.07	82.935 8	0.02
Ionosphere	0.024 3	0.379 5	6.39	43.882 4	0.06	150.410 6	0.02
Wdbc	0.060 9	0.844 2	7.22	32.565 8	0.19	109.570 9	0.06
German	0.181 8	2.704 0	6.72	167.698 3	0.11	97.490 2	0.19
Waveform	5.971 3	104.561 1	5.71	—	—	—	—

Fuzzy-rough set based unsupervised dynamic feature selection algorithm

基于模糊粗糙集的无监督动态特征选择算法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 8

References 22

Related Articles 15

Recommended Articles

Metrics

数据集	UDFRFS		UFRFS		FRUAR		SUIFR
数据集	基础	增量	基础	增量	基础	增量	基础	增量
Olive Oil	4.74	4.81	4.74	4.81	1.10	1.00	1.00	1.00
SCADI	17.10	17.36	17.10	17.36	16.94	16.47	1.00	1.00
SAP	7.77	9.17	7.77	9.17	7.88	9.77	1.00	1.00
Divorce	9.89	10.23	9.89	10.23	10.03	10.47	1.00	1.00
Sonar	8.14	8.14	8.14	8.14	4.76	3.91	3.06	3.00
Ionosphere	10.06	11.10	10.06	11.10	12.00	14.15	12.48	14.62
Wdbc	7.94	8.54	7.94	8.54	5.20	4.68	3.31	3.00
German	9.46	11.08	9.46	11.08	9.47	11.15	1.00	1.00
Waveform	9.22	11.98	9.22	11.98	—	—

[1]	Guixiang XUE, Hui WANG, Weifeng ZHOU, Yu LIU, Yan LI. Port traffic flow prediction based on knowledge graph and spatio-temporal diffusion graph convolutional network [J]. Journal of Computer Applications, 2024, 44(9): 2952-2957.
[2]	Hong CHEN, Bing QI, Haibo JIN, Cong WU, Li’ang ZHANG. Class-imbalanced traffic abnormal detection based on 1D-CNN and BiGRU [J]. Journal of Computer Applications, 2024, 44(8): 2493-2499.
[3]	Shuo SUN, Wei ZHANG, Wendi FENG, Yuwei ZHANG. Automatic foreign function interface generation method based on source code analysis [J]. Journal of Computer Applications, 2024, 44(7): 2151-2159.
[4]	Tianci KE, Jianhua LIU, Shuihua SUN, Zhixiong ZHENG, Zijie CAI. Aspect-level sentiment analysis model combining strong association dependency and concise syntax [J]. Journal of Computer Applications, 2024, 44(6): 1786-1795.
[5]	Lin GAO, Yu ZHOU, Tak Wu KWONG. Evolutionary bi-level adaptive local feature selection [J]. Journal of Computer Applications, 2024, 44(5): 1408-1414.
[6]	Mingzhu LEI, Hao WANG, Rong JIA, Lin BAI, Xiaoying PAN. Oversampling algorithm based on synthesizing minority class samples using relationship between features [J]. Journal of Computer Applications, 2024, 44(5): 1428-1436.
[7]	Zihan LIU, Dengwen ZHOU, Yukai LIU. Image super-resolution network based on global dependency Transformer [J]. Journal of Computer Applications, 2024, 44(5): 1588-1596.
[8]	Lin SUN, Menghan LIU. K-means clustering based on adaptive cuckoo optimization feature selection [J]. Journal of Computer Applications, 2024, 44(3): 831-841.
[9]	Dapeng XU, Xinmin HOU. Feature selection method for graph neural network based on network architecture design [J]. Journal of Computer Applications, 2024, 44(3): 663-670.
[10]	Shengjie MENG, Wanjun YU, Ying CHEN. Feature selection algorithm for high-dimensional data with maximum correlation and maximum difference [J]. Journal of Computer Applications, 2024, 44(3): 767-771.
[11]	Jingxin LIU, Wenjing HUANG, Liangsheng XU, Chong HUANG, Jiansheng WU. Unsupervised feature selection model with dictionary learning and sample correlation preservation [J]. Journal of Computer Applications, 2024, 44(12): 3766-3775.
[12]	Yanbo LI, Qing HE, Shunyi LU. Aspect sentiment triplet extraction integrating semantic and syntactic information [J]. Journal of Computer Applications, 2024, 44(10): 3275-3280.
[13]	Wanting JI, Wenyi LU, Yuhang MA, Linlin DING, Baoyan SONG, Haolin ZHANG. Machine reading comprehension event detection based on relation-enhanced graph convolutional network [J]. Journal of Computer Applications, 2024, 44(10): 3288-3293.
[14]	Tian HE, Zongxin SHEN, Qianqian HUANG, Yanyong HUANG. Adaptive learning-based multi-view unsupervised feature selection method [J]. Journal of Computer Applications, 2023, 43(9): 2657-2664.
[15]	Hongjun HENG, Dingcheng YANG. Knowledge enhanced aspect word interactive graph neural network [J]. Journal of Computer Applications, 2023, 43(8): 2412-2419.

数据集	UDFRFS		UFRFS		FRUAR		SUIFR		RAW1
数据集	基础	增量	基础	增量	基础	增量	基础	增量	基础	增量
Olive Oil	71.9	74.3	71.9	74.3	42.6	44.2	52.7	42.8	83.3	83.3
SCADI	73.7	74.4	73.7	74.4	77.1	76.2	45.1	45.0	80.8	80.8
SAP	43.1	43.3	43.1	43.3	43.5	43.4	42.9	41.8	43.7	42.1
Divorce	97.7	97.7	97.7	97.7	97.7	97.8	91.7	89.1	97.6	97.7
Sonar	78.0	76.9	78.0	76.9	61.5	58.7	60.4	62.2	83.8	85.3
Ionosphere	86.7	87.5	86.7	87.5	86.1	86.3	87.0	86.5	85.8	85.4
Wdbc	94.3	94.6	94.3	94.6	85.9	83.6	76.0	72.0	95.3	96.2
German	90.5	90.2	90.5	90.2	88.7	90.0	80.0	79.4	69.1	70.6
Waveform	72.0	76.2	72.0	76.2	—	—			73.1	78.2

数据集	UDFRFS		UFRFS		FRUAR		SUIFR		RAW1
数据集	基础	增量	基础	增量	基础	增量	基础	增量	基础	增量
Olive Oil	71.9	74.3	71.9	74.3	42.6	44.2	52.7	42.8	83.3	83.3
SCADI	73.7	74.4	73.7	74.4	77.1	76.2	45.1	45.0	80.8	80.8
SAP	43.1	43.3	43.1	43.3	43.5	43.4	42.9	41.8	43.7	42.1
Divorce	97.7	97.7	97.7	97.7	97.7	97.8	91.7	89.1	97.6	97.7
Sonar	78.0	76.9	78.0	76.9	61.5	58.7	60.4	62.2	83.8	85.3
Ionosphere	86.7	87.5	86.7	87.5	86.1	86.3	87.0	86.5	85.8	85.4
Wdbc	94.3	94.6	94.3	94.6	85.9	83.6	76.0	72.0	95.3	96.2
German	90.5	90.2	90.5	90.2	88.7	90.0	80.0	79.4	69.1	70.6
Waveform	72.0	76.2	72.0	76.2	—	—			73.1	78.2

数据集	UDFRFS		UFRFS		FRUAR		SUIFR		RAW1
数据集	基础	增量	基础	增量	基础	增量	基础	增量	基础	增量
Olive Oil	73.3	76.7	73.3	76.7	45.1	47.1	56.6	52.1	24.1	4.3
SCADI	74.3	75.5	74.3	75.5	78.1	77.0	44.8	44.3	81.5	82.9
SAP	45.8	45.2	45.8	45.2	45.6	45.5	40.9	39.0	46.5	47.7
Divorce	97.7	97.8	97.7	97.8	97.5	97.5	91.5	88.8	97.5	97.2
Sonar	67.3	65.5	67.3	65.5	59.2	58.7	60.7	60.9	67.7	68.5
Ionosphere	81.0	82.3	81.0	82.3	81.9	83.3	84.9	84.2	81.3	81.9
Wdbc	92.6	92.6	92.6	92.6	82.9	80.2	75.6	70.3	92.7	93.1
German	87.7	88.6	87.7	88.6	85.9	88.2	77.6	79.3	72.2	73.1
Waveform	74.3	78.2	74.3	78.2	—	—	—	—	75.5	79.8

数据集	UDFRFS		UFRFS		FRUAR		SUIFR		RAW1
数据集	基础	增量	基础	增量	基础	增量	基础	增量	基础	增量
Olive Oil	73.3	76.7	73.3	76.7	45.1	47.1	56.6	52.1	24.1	4.3
SCADI	74.3	75.5	74.3	75.5	78.1	77.0	44.8	44.3	81.5	82.9
SAP	45.8	45.2	45.8	45.2	45.6	45.5	40.9	39.0	46.5	47.7
Divorce	97.7	97.8	97.7	97.8	97.5	97.5	91.5	88.8	97.5	97.2
Sonar	67.3	65.5	67.3	65.5	59.2	58.7	60.7	60.9	67.7	68.5
Ionosphere	81.0	82.3	81.0	82.3	81.9	83.3	84.9	84.2	81.3	81.9
Wdbc	92.6	92.6	92.6	92.6	82.9	80.2	75.6	70.3	92.7	93.1
German	87.7	88.6	87.7	88.6	85.9	88.2	77.6	79.3	72.2	73.1
Waveform	74.3	78.2	74.3	78.2	—	—	—	—	75.5	79.8