摘要
提出了一种基于太赫兹(THz)光谱技术以及布谷鸟搜索(CS)算法优化支持向量机(SVM)的有效的转基因产品鉴别方法(CS-SVM)。实验采用太赫兹时域光谱(THz-TDS)系统测量了三种转基因大豆种子及其亲本样品在0.2~1.2THz波段的THz光谱,并采用SVM方法对转基因和非转基因大豆种子进行了分类鉴别研究,其中SVM的两个重要参数(惩罚因子和核参数)采用CS算法进行优化。实验结果表明,应用THz光谱技术结合CS-SVM方法为转基因和非转基因生物的检测和识别提供了一种快速、无损和可靠的分析方法。
This paper develops an effective identification method to discriminate genetically modified(GM)and non-GM organisms.The method is proposed based on terahertz(THz)spectroscopy and support vector machines optimized by Cuckoo Search algorithm(CS-SVM).In this study,the THz spectra of three GM and non-GM soya seed samples were obtained by using terahertz time-domain spectroscopy(THz-TDS)system between 0.2and 1.2THz.Then,the SVM model is employed to distinguish GM and non-GM soya seeds,in which the two crucial parameters,including the penalty factor and kernel parameter,are optimized by CS algorithm.The experimental results show that THz spectroscopy combined with CS-SVM can provide a rapid,reliable and non-invasive method for GMOs and non-GMOs discrimination.
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2017年第2期618-623,共6页
Spectroscopy and Spectral Analysis
基金
the National Natural Science Foundation of China(11574059)
partly supported by the Guangxi Natural Science Foundation(2015GXNSFBA139252)
partly supported by the program(YQ15104)from Guangxi Key Laboratory of Automatic Detecting Technology and Instruments
关键词
太赫兹光谱
转基因生物
支持向量机
布谷鸟搜索算法
Terahertz spectroscopy
Genetically modified organisms
Support vector machines
Cuckoo Search algorithm