[1]赵建印,孙敏奇,崔 洋,等.小样本离散时序数据的集成学习特征提取方法[J].机械与电子,2026,44(07):13-22.
　ZHAO Jianyin,SUN Minqi,CUI Yang,et al.An Ensemble Learning-driven Feature Extraction Method for Small-sample Discrete Time-series Data[J].Machinery & Electronics,2026,44(07):13-22.
点击复制

小样本离散时序数据的集成学习特征提取方法()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
44
期数:
2026年07期
页码:
13-22
栏目:
研究与设计
出版日期:
2026-07-25

文章信息/Info

Title:
An Ensemble Learning-driven Feature Extraction Method for Small-sample Discrete Time-series Data
文章编号:
1001-2257(2026)07-0013-10
作者:
赵建印; 孙敏奇; 崔 洋; 秦玉峰
海军航空大学,山东 烟台 264001
Author(s):
ZHAO Jianyin; SUN Minqi; CUI Yang; QIN Yufeng
(Naval Aviation University,Yantai 264001,China)
关键词:
Mann-Kendall趋势检验; 特征提取; 核主成分分析; 偏离度; 离散时序数据
Keywords:
Mann Kendall trend test; feature extraction; kernel principal component analysis; deviation degree; discrete time-series data
分类号:
TP18;TP391.4
文献标志码:
A
摘要:
针对高精度设备多部件、多模式退化、长检测周期下的高维小样本时序数据特征提取难题,提出一种融合改进Mann-Kendall(MK)与核主成分分析(KPCA)的特征提取方法。该方法先对时序检测数据进行标准化处理,针对传统MK 算法无法识别多模式退化特征、无冗余筛选能力的问题,构建滑动窗口MK趋势分析与加权欧氏距离偏离度分析模型,实现数值趋势型、方差增大型退化特征的精准识别与冗余特征的高效精简;再针对传统KPCA 核参数选择盲目、直接降维丢失耦合信息的问题,构建核参数自适应调优与分步降维的改进KPCA 模型,基于筛选后的核心退化特征精准建模非线性耦合关系,最终融合生成一维健康指标(HI)。实验结果表明,所提方法在小样本适应性、特征筛选精度及非线性耦合建模能力上显著优于传统方法,可为复杂机电设备的健康状态评估提供有效技术支撑。
Abstract:
Accurately extracting features from high-dimensional,small sample,time-series data remains a critical challenge in the health assessment of high-precision equipment,which typically exhibits multi-component configurations,multi-mode degradation behaviors,and long inspection intervals.To address this problem,this paper proposes a feature extraction method that integrates an improved Mann-Kendall (MK) test with kernel principal component analysis (KPCA).After standardizing the raw timeseries measurements,we first construct a sliding window MK trend analysis model coupled with a weighted Euclidean distance deviation analysis module.This design overcomes the inherent limitations of the conventional MK algorithm,namely,its inability to recognize multi-mode degradation patterns and its lack of redundancy filtering capability—thereby enabling precise identification of both monotonic trend type and variance increment type degradation features while efficiently eliminating redundant variables.Subsequently, to circumvent the arbitrary selection of kernel parameters and the loss of coupling information inherent in direct dimensionality reduction using standard KPCA,we develop an improved KPCA framework with adaptive kernel parameter optimization and a two-stage dimensionality reduction strategy.This framework builds a nonlinear coupling model based on the selected core degradation features,and finallyfuses them into a one dimensional health indicator (HI).Experimental results demonstrate that the proposed method significantly outperforms traditional methods in terms of adaptability to small sample scenarios, feature selection accuracy,and nonlinear coupling modeling capability,providing effective technical support for health condition assessment of complex electromechanical equipment.

参考文献/References:

[1] 陈凯诺,张福光,韩建立,等.基于Transformer小样本多源数据融合的装备剩余寿命预测评估[J].装备环境工程,2024,21(11):65-73.

[2] 赵建印,姜经纬,魏顺涛.基于改进无迹卡尔曼滤波的可靠性评估方法[J].现代防御技术,2025,53(4):129-139.
[3] 汪攀,刘毅敏.Sen’s斜率估计与Mann-Kendall法在设备运行趋势分析中的应用[J].武汉科技大学学报,2014,37(6):454-457.
[4] 岑潮宇,代亮成,池茂儒,等.基于KPCA-SO-KELM的抗蛇行减振器故障诊断[J].科学技术与工程,2025,25(11):4551-4558.
[5] 马天池,沈君贤,宋狄,等.基于多域特征与信息融合的叶片裂纹故障诊断[J].东南大学学报(自然科学版),2024,54(6):1567-1573.
[6] Then Y J,Halim S A.Modified Mann-Kendall with higher-order statistics for trend analysis[J].Scientific Reports,2026,16:500.
[7] Kaib M T H,Kouadri A,Harkat M F,et al.Improvement of kernel principal component analysis-based approach for nonlinear process monitoring by data set size reduction using class interval[J].IEEE Access,2024,12:11470-11480.
[8] Anowar F,Sadaoui S,Selim B.Conceptual and empirical comparison of dimensionality reduction algorithms (PCA,KPCA,LDA,MDS,SVD,LLE,ISOMAP,LE, ICA,t-SNE)[J].Computer Science Review,2021,40:100378.
[9] Chen Chuang,Li Mengchen,Shi Jiantao,et al.Constructing high-quality health indicators from multi-source sensor data for predictive maintenance applications[ J].Expert Systems with Applications,2025,285:127870.
[10] 安华,王国锋,王喆,等.基于深度学习理论的刀具状态监测及剩余寿命预测方法[J].电子测量与仪器学报,2019,33(9):64-70.
[11] Huang Faming,Zhang Jing,Zhou Chuangbing,et al. A deep learning algorithm using a fully connected sparse autoencoder neural network for landslide susceptibility prediction [J].Landslides,2020,17:217-229.
[12] Lee S,Kim H J,Kim S B.Dynamic dispatching system using a deep denoising autoencoder for semiconductor manufacturing[J].Applied Soft Computing,2020,86:105904.
[13] Balevi E,Andrews J G.Autoencoder-based error correction coding for one-bit quantization[J].IEEE Transactions on Communications,2020,68 (6): 3440-3451.
[14] Elsheikh A,Yacout S,Ouali M.Bidirectional handshaking LSTM for remaining useful life prediction[J].Neurocomputing,2019,323:148-156.
[15] Yu Yong,Hu Changhua,Si Xiaosheng,et al.Averaged Bi-LSTM networks for RUL prognostics with non life-cycle labeled dataset[J].Neurocomputing,2020, 402:134-147.
[16] Yang Boyuan,Liu Ruonan,Zio E.Remaining useful life prediction based on a double-convolutional neural network architecture[J].IEEE Transactions on Industrial Electronics,2019,66(12):9521-9530.
[17] Zbontar J,Jing Li,MISRA I,et al.Barlow twins:Self-supervised learning via redundancy reduction[C]∥Proceedings of the 38th International Conference on Machine Learning (ICML 2021),2021:12310-12320.
[18] Si Lei,Wang Zhongbin,Tan Chao,et al.A feature extraction method-based on composite multi-scale permutation entropy and Laplacian score for shearer cutting state recognition[J].Measurement,2019,145:84-93.
[19] Liu Han,Zhou Jianzhong,Zheng Yang,et al.Fault diagnosis of rolling bearings with recurrent neural network-based autoencoders[J].ISA Transactions, 2018,77:167-178.

相似文献/References:

[1]孙 佩,胡小敏.基于图像特征提取的地铁客流量自动统计系统[J].机械与电子,2020,(10):17.
　SUN Pei,HU Xiaomin.Metro Passenger Flow Statistics System Based on Image Feature Extraction[J].Machinery & Electronics,2020,(07):17.
[2]韦永奇,胥光申,刘 晖,等.基于图像识别的激光切割质量分析[J].机械与电子,2021,(01):56.
　WEI Yongqi,XU Guangshen,LIU Hui,et al.Laser Cutting Quality Analysis Based on Image Recognition[J].Machinery & Electronics,2021,(07):56.
[3]庄爱军.基于WA-ESN的建筑起重机械故障检测[J].机械与电子,2021,(01):67.
　ZHUANG Aijun.Fault Detection of Construction Crane Based on WA-ESN[J].Machinery & Electronics,2021,(07):67.
[4]杨玥旻,闫维新.基于图像处理的爬壁机器人焊缝识别与跟踪[J].机械与电子,2021,(03):65.
　YANG Yuemin,YAN Weixin.Welding Seam Recognition and Tracking Based on Image Processing for Wall-climbing Robot[J].Machinery & Electronics,2021,(07):65.
[5]许 哲,张少帅,郭 璐,等.无人机深度学习去雾算法[J].机械与电子,2021,(04):13.
　XU Zhe,ZHANG Shaoshuai,GUO Lu,et al. Deep Learning Defogging Algorithm for UAV[J].Machinery & Electronics,2021,(07):13.
[6]王维坤,余志伟,马鹏飞.输电线路巡检智能单兵装备全景采集系统设计[J].机械与电子,2021,(09):32.
　WANG Weikun,YU Zhiwei,MA Pengfei.Design of Panoramic Acquisition System for Intelligent Single Soldier Equipment in Transmission Line Inspection[J].Machinery & Electronics,2021,(07):32.
[7]冯 洋,尹 松,闫敬东,等.基于数字孪生的电网变压器故障诊断方法[J].机械与电子,2022,(06):26.
　FENG Yang,YIN Song,YAN Jingdong,et al.Fault Diagnosis Method of Power Grid Transformer Based on Digital Twin[J].Machinery & Electronics,2022,(07):26.
[8]高正贤,王 靳,王建永.基于 SVDI 的低功耗自供电机械设备故障检测方法[J].机械与电子,2023,41(02):8.
　GAO Zhengxian,WANG Jin,WANG Jianyong.Fault Detection Method for Low-power Self-powered Mechanical Equipment Based on SVDI[J].Machinery & Electronics,2023,41(07):8.
[9]权 超,穆龙涛.基于多 SAE 的机械臂运行可靠性深度特征提取与融合[J].机械与电子,2023,41(06):15.
　QUAN Chao,MU Longtao.Reliability Depth Feature Extraction and Fusion of Robotic Arm Based on Multi-SAE[J].Machinery & Electronics,2023,41(07):15.
[10]郭玉龙,陈一兵.基于线阵相机的煤矿机械加工表面质量检测方法[J].机械与电子,2023,41(06):21.
　GUO Yulong,CHEN Yibing.Detection Method of Surface Quality of Coal Mine Machining Based on Line Scan Camera[J].Machinery & Electronics,2023,41(07):21.

备注/Memo

备注/Memo:
收稿日期:2026-05-22
作者简介:赵建印 (1976-),男,河北衡水人,博士,教授,研究方向为装备可靠性;秦玉峰 (1995-),男,黑龙江大庆人,博士,讲师,研究方向为装备可靠性,通信作者,E mail:Hy_qyf082@163.com。
更新日期/Last Update: 2026-08-27