基于模糊隶属度的近红外光谱模型鲁棒性分析

被引:1
作者
高珏 [1 ,2 ,3 ]
李海森 [1 ,2 ]
徐超 [1 ,2 ]
朱培逸 [3 ]
机构
[1] 哈尔滨工程大学水声工程学院
[2] 哈尔滨工程大学水声技术重点实验室
[3] 常熟理工学院电气与自动化工程学院
关键词
鲁棒性; 模糊隶属度; 近红外光谱; 建模; 噪声; 数据域描述;
D O I
暂无
中图分类号
O434.3 [红外线]; O212.1 [一般数理统计];
学科分类号
070103 [概率论与数理统计]; 070207 [光学];
摘要
针对近红外光谱模型存在的鲁棒性问题,在模型建立时引入模糊隶属度,提出了一种自动生成模糊隶属度的方法。建立光谱样本的数据域描述函数,引入信任因子和舍弃因子,通过映射关系得到模糊隶属度函数,参数寻优后自动生成每个样本的模糊隶属度。在此基础上建立了基于FSVM的苹果糖度回归模型。试验结果表明,对比常规的MLR、PLSR和SVM模型,FSVM模型在训练样本变化和高斯噪声、乘性噪声、基线漂移、基线倾斜和波长漂移这5种噪声的分别作用下表现出最佳的性能。模糊隶属度的引入提高了近红外光谱模型的泛化能力和抗噪能力,改善了模型的鲁棒性。
引用
收藏
页码:312 / 316
页数:5
相关论文
共 17 条
[1]
A Review on the Use of Near-Infrared Spectroscopy for Analyzing Feed Protein Materials [J].
Chen, Longjian ;
Yang, Zengling ;
Han, Lujia .
APPLIED SPECTROSCOPY REVIEWS, 2013, 48 (07) :509-522
[2]
Fuzzy support vector machine based on within-class scatter for classification problems with outliers or noises.[J].Wenjuan An;Mangui Liang.Neurocomputing.2013,
[3]
Robust classifier learning with fuzzy class labels for large-margin support vector machines.[J].Chan-Yun Yang;Jui-Jen Chou;Feng-Li Lian.Neurocomputing.2013,
[5]
Estimating olive leaf nitrogen concentration using visible and near-infrared spectral reflectance.[J].Nativ Rotbart;Zeev Schmilovitch;Yafit Cohen;Victor Alcahantis;Ran Erel;Timea Ignat;Clara Shenderey;Arnon Dag;Uri Yermiyahu.Biosystems Engineering.2012,
[6]
Near-infrared (NIR) and mid-infrared (MIR) spectroscopic techniques for assessing the amount of carbon stock in soils – Critical review and research perspectives.[J].Véronique Bellon-Maurel;Alex McBratney.Soil Biology and Biochemistry.2011, 7
[7]
How to build a robust model against perturbation factors with only a few reference values: A chemometric challenge at ‘Chimiométrie 2007’.[J].Juan Antonio Fernández Pierna;Fabian Chauchard;Sébastien Preys;Jean Michel Roger;Oswin Galtier;Vincent Baeten;Pierre Dardenne.Chemometrics and Intelligent Laboratory Systems.2010, 2
[8]
Postharvest quality of apple predicted by NIR-spectroscopy: Study of the effect of biological variability on spectra and model performance.[J].Els Bobelyn;Anca-Sabina Serban;Mihai Nicu;Jeroen Lammertyn;Bart M. Nicolai;Wouter Saeys.Postharvest Biology and Technology.2009, 3
[9]
Nondestructive measurement of fruit and vegetable quality by means of NIR spectroscopy:: A review [J].
Nicolai, Bart M. ;
Beullens, Katrien ;
Bobelyn, Els ;
Peirs, Ann ;
Saeys, Wouter ;
Theron, Karen I. ;
Lammertyn, Jeroen .
POSTHARVEST BIOLOGY AND TECHNOLOGY, 2007, 46 (02) :99-118
[10]
Support vector data description [J].
Tax, DMJ ;
Duin, RPW .
MACHINE LEARNING, 2004, 54 (01) :45-66