基于傅里叶近红外特征光谱的血流感染致病菌鉴别研究

发布时间：2018-02-22 20:54

  本文关键词： 血流感染 傅里叶变换近红外光谱 偏最小二乘判别分析 最小二乘-支持向量机 竞争性自适应重加权算法 病原菌鉴别　出处：《福州大学学报(自然科学版)》2017年05期 　论文类型：期刊论文

【摘要】：利用傅里叶变换近红外光谱(FT-NIR)收集1 000~1 852 nm范围内3种常见病原菌大肠杆菌(ATCC25922)、金黄色葡萄球菌(ATCC 29213)、铜绿假单胞菌(ATCC 27853)的近红外透射光谱,采用竞争性自适应重加权算法(CARS)对波长变量进行筛选,并分别结合偏最小二乘判别分析(PLS-DA)、最小二乘-支持向量机(LS-SVM)建立鉴别模型.比较两种鉴别模型在进行波长变量优选前后的性能发现,采用全波段建模的PLS-DA与LS-SVM两种模型的预测性能较低;利用CARS对波长变量进行筛选后,对优选的24个特征波长分别建立两种鉴别模型,模型预测性能明显提高,其中以LS-SVM模型最优,3种病原菌准确率分别为85.0%,100%和100%.研究结果表明,利用CARS能够有效去除光谱无用信息,减少模型复杂度,增强模型预测性能,结合LS-SVM可为临床利用近红外快速检测血流感染病原菌提供一种新的方法.
[Abstract]:The near infrared transmission spectra of Escherichia coli ATCC25922, Staphylococcus aureus ATCC29213and Pseudomonas aeruginosa ATCC27853 were collected by Fourier transform near infrared spectroscopy (FT-NIR). A competitive adaptive reweighting algorithm (CARSs) is used to screen the wavelength variables. Combined with partial least squares discriminant analysis (PLS-DAA) and least squares support vector machine (LS-SVM), the identification models were established, and the performance of the two discriminant models before and after optimal selection of wavelength variables were compared. The prediction performance of PLS-DA and LS-SVM models based on full-band modeling is low, and the prediction of the two models can be improved obviously by using CARS to screen the wavelength variables and to establish two discriminant models for the 24 characteristic wavelengths selected separately. The accuracy of LS-SVM model was 85.0% and 100%, respectively. The results showed that CARS could effectively remove spectral useless information, reduce the complexity of model, and enhance the performance of model prediction. The combination of LS-SVM can provide a new method for rapid detection of blood stream infection pathogens by near infrared spectroscopy.
【作者单位】： 福州大学电气工程与自动化学院;福建省医疗器械和医药技术重点实验室;福建医科大学医学技术与工程学院;
【基金】：国家自然科学基金资助项目(61403319) 福建省科技厅国际合作资助项目(2015I003) 福建省教育厅科技资助项目(JK2014001)
【分类号】：O657.33;R446.5
，

本文编号：1525285