烟用爆珠内液质量稳定性检测——基于紫外光谱技术结合SVR算法
|
摘要
【目的】基于紫外(UV)光谱技术和支持向量机回归(SVR)算法,建立一种烟用爆珠内液判别和稳定性分析方法。【方法】采用无水乙醇稀释爆珠内液,UV扫描,光谱经预处理后,建立四种不同牌号烟用爆珠SVR模型,并对模型进行了验证。【结果】(1)最佳UV图谱预处理方法是平滑后归一化;最佳SVR类型是ν-SVR;最佳核函数是径向基核函数。(2)四个牌号烟用爆珠内液ν-SVR模型校正集分类变量的预测值与实测值的相关系数均≥0.9993,SVR模型对于参与建模的400个烟用爆珠内液样本(校正集)和未参与建模的80个烟用爆珠内液样本(验证集)的预测准确度均为100%,模型拟合性好,预测精度高,判别能力强。(3)基于SVR模型分类变量值建立的单值控制图可以对烟用爆珠内液的稳定性进行快速判定,判定结果与GC/MS检测结果一致。【结论】紫外可见光谱技术结合SVR算法可对不同牌号烟用爆珠内液质量进行有效判别,且方法快速准确、经济环保、易于推广。
Based on ultraviolet(UV) spectroscopy and SVR(support vector machine regression) algorithm, a method for the determination and stability analysis of breakable capsules in cigarette was established. The internal liquid of breakable capsules was diluted with anhydrous ethanol, followed by UV scanning and spectral pretreatment, four SVR models of different capsule brands were established and validated. Results show that:(1) The optimal UV spectrum pretreatment method was smoothing and normalization, the optimal type of SVR was ν-SVR, and the optimal kernel function was radial basis function.(2) The correlation coefficient between predicted and measured values of the classification variables in the ν-SVR model calibration set were all more than 0.9993 for four capsule brands. The prediction accuracy of SVR model was 100% for both 400 samples of breakable capsule inner liquid(calibration set) and 80 samples of breakable capsule inner liquid(validation set) that were not involved in the modeling, indicating high accuracy and efficiency.(3) The stability of breakable capsules could be effectively evaluated by quality control chart based on classification variable values of SVR model, and the evaluation results were consistent with those of GC/MS method. It can be concluded that UV spectra combined with SVR algorithm could be used for quality evaluation of liquid in cigarette breakable capsules, and the method was rapid, accurate, economical, and environmentfriendly.
Based on ultraviolet(UV) spectroscopy and SVR(support vector machine regression) algorithm, a method for the determination and stability analysis of breakable capsules in cigarette was established. The internal liquid of breakable capsules was diluted with anhydrous ethanol, followed by UV scanning and spectral pretreatment, four SVR models of different capsule brands were established and validated. Results show that:(1) The optimal UV spectrum pretreatment method was smoothing and normalization, the optimal type of SVR was ν-SVR, and the optimal kernel function was radial basis function.(2) The correlation coefficient between predicted and measured values of the classification variables in the ν-SVR model calibration set were all more than 0.9993 for four capsule brands. The prediction accuracy of SVR model was 100% for both 400 samples of breakable capsule inner liquid(calibration set) and 80 samples of breakable capsule inner liquid(validation set) that were not involved in the modeling, indicating high accuracy and efficiency.(3) The stability of breakable capsules could be effectively evaluated by quality control chart based on classification variable values of SVR model, and the evaluation results were consistent with those of GC/MS method. It can be concluded that UV spectra combined with SVR algorithm could be used for quality evaluation of liquid in cigarette breakable capsules, and the method was rapid, accurate, economical, and environmentfriendly.
引文
[1]何媛,彭军仓,吕娟,等.GC/MS指纹图谱及系统聚类分析用于烟用爆珠内含物的测定[J].烟草科技,2018,51(7):54-60.He Yuan,Peng Juncang,Lu Juan,et al.GC/MS fingerprinting with hierarchical cluster analysis to determine components in breakable capsules in cigarette filter[J].Tobacco Science&Technology,2018,51(7):54-60.
[2]朱立军,李宁.一种烟用香珠粒径大小的快速检验方法:CN103884631A[P].2014-06-25.Zhu Lijun,Li Ning.A rapid method for the size of breakable capsules in cigarette filter:CN 103884631A[P].2014-06-25.
[3]刘秀明,张健,刘亚,等.高效液相色谱法同时测定爆珠壁材中8种水溶性着色剂[J].中国测试,2018,44(3):48-52.Liu Xiuming,Zhang Jian,Liu Ya,et al.Simultaneous determination of 8 water soluble colorants in wall materials of cigarette blasting bead by high performance liquid chromatography[J].China Measurement&Test,2018,44(3):48-52.
[4]朱亚峰,胡军,唐荣成,等.卷烟滤嘴加香研究进展[J].中国烟草学报,2011,17(6):104-109.Zhu Yafeng,Hu Jun,Tang Rongcheng,et al.Recent advances in research and development in cigarette filter flavoring technique[J].Acta Tabacaria Sinica,2011,17(6):104-109.
[5]朴洪伟,金勇华,金钟国,等.甜橙香胶囊滤棒对烟气有害成分及卷烟香气特性的影响[J].郑州轻工业学院学报(自然科学版),2015,30(3/4):48-51.Piao Hongwei,Jin Yonghua,Jin Zhongguo,et al.Effect of filter containing sweet-orange flavor capsule on harmful components in cigarette smoke and cigarette flavor characteristics[J].Journal of Zhengzhou University of Light Industry(Natural Science),2015,30(3/4):48-51.
[6]朱风鹏,李雪,罗彦波,等.爆珠破碎对主流烟气有害成分释放量和滤嘴截留的影响[J].烟草科技,2017,50(4):37-41.Zhu Fengpeng,Li Xue,Luo Yanbo,et al.Effects of capsule crushing on cigarette filter’s filtration and delivery of seven harmful components in mainstream cigarette smoke[J].Tobacco Science&Technology,2017,50(4):37-41.
[7]张晋,张琳.一种爆珠滤棒中爆珠的微波测定方法:CN105866140A[P].2016-08-17.Zhang Jin,Zhang Lin.A microwave method for capsule in capsule filter:CN105866140A[P].2016-08-17.
[8]瞿海斌,刘晓宣,程翼宇.中药材三七提取液近红外光谱的支持向量机回归校正方法[J].高等学校化学学报,2004,25(1):39-43.Qu Haibin,Liu Xiaoxuan,Cheng Yiyu.A method for analysis of extracting solution of Panax Notoginseng using near-infrared spectroscopy and support vector regression[J].Chemical Journal of Chinese Universities,2004,25(1):39-43.
[9]王国祥,王海燕,王虎,等.基于Raman光谱和支持向量机回归的古井贡酒年份鉴别方法[J].光谱学与光谱分析,2016,36(3):729-735.Wang Guoxiang,Wang Haiyan,Wang Hu,et al.Study on recognition of liquor age of Gujing based on Raman spectra and support vector regression[J].Spectroscopy and Spectral Analysis,2016,36(3):729-735.
[10]熊宇虹,荣祺,舒明磊,等.改进的基于线性支持向量机回归的近红外光谱分析方法[J].上海应用技术学院学报(自然科学版),2016,16(3):275-279..Xiong Yuhong,Rong Qi,Shu Minglei,et al.Improved near infrared spectral analysis method based on linear support vector machine regression[J].Journal of Shanghai Institute of Technology(Natural Science),2016,16(3):275-279.
[11]黎洪利,赵朋贤,杨俊,等.紫外指纹图谱在烟用香精质量控制中的应用[J].中国烟草学报,2009,15(4):19-21.Li Hongli,Zhao Pengxian,Yang Jun,et al.Application of ultraviolet fingerprint spectrum in tobacco flavor quality control[J].Acta Tabacaria Sinica,2009,15(4):19-21.
[12]孙志涛.紫外-可见分光光度法在烟用香精香料质量评价中的应用[J].安徽农学通报,2010,16(2):144-147.Sun Zhitao.Application of ultraviolet-visible spectrum piin tobacco flavor quality evaluation[J].Anhui Agricultural Science Bulletin,2010,16(2):144-147.
[13]余志雄,周创兵,李俊平,等.基于ν-SVR算法的边坡稳定性预测[J].岩石力学与工程学报,2005,24(14):2468-2475.Yu Zhixiong,Zhou Chuangbing,Li Junping,et al.Predicting slope stability based onν-SVR algorithm[J].Chinese Journal of Rock Mechanics and Engineering,2005,24(14):2468-2475.
[14]Chang Chih-Chung,Lin Chih-Jen.Trainingν-Support Vector Regression:Theory and Algorithms[J].Neural Computation,2002,14(8):1959-1977.
[15]Huang Bo,Cai Zhihua,Gu Qiong,et al.Using Support Vector Regression for Classification[C].International Conference on Advanced Data Mining and Applications,2008,5139:581-588.
[16]Chen Muyen,Chen Chiachen,Chang Yafen.Comparison of support vector machine and support vector regression:An application to predict financial distress and bankruptcy[C].International Conference on Service Systems&Service Management,2010,53(5):1-6.
[17]吴元清,杜树新,严赟,等.水体有机污染物浓度检测中的紫外光谱分析方法[J].光谱学与光谱分析,2011,31(1):233-237.Wu Yuanqing,Du Shuxin,Yan Yun,et al.Ultraviolet spectrum analysis methods for detecting the concentration of organic pollutants in water[J].Spectroscopy and Spectral Analysis,2011,31(1):233-237.
[2]朱立军,李宁.一种烟用香珠粒径大小的快速检验方法:CN103884631A[P].2014-06-25.Zhu Lijun,Li Ning.A rapid method for the size of breakable capsules in cigarette filter:CN 103884631A[P].2014-06-25.
[3]刘秀明,张健,刘亚,等.高效液相色谱法同时测定爆珠壁材中8种水溶性着色剂[J].中国测试,2018,44(3):48-52.Liu Xiuming,Zhang Jian,Liu Ya,et al.Simultaneous determination of 8 water soluble colorants in wall materials of cigarette blasting bead by high performance liquid chromatography[J].China Measurement&Test,2018,44(3):48-52.
[4]朱亚峰,胡军,唐荣成,等.卷烟滤嘴加香研究进展[J].中国烟草学报,2011,17(6):104-109.Zhu Yafeng,Hu Jun,Tang Rongcheng,et al.Recent advances in research and development in cigarette filter flavoring technique[J].Acta Tabacaria Sinica,2011,17(6):104-109.
[5]朴洪伟,金勇华,金钟国,等.甜橙香胶囊滤棒对烟气有害成分及卷烟香气特性的影响[J].郑州轻工业学院学报(自然科学版),2015,30(3/4):48-51.Piao Hongwei,Jin Yonghua,Jin Zhongguo,et al.Effect of filter containing sweet-orange flavor capsule on harmful components in cigarette smoke and cigarette flavor characteristics[J].Journal of Zhengzhou University of Light Industry(Natural Science),2015,30(3/4):48-51.
[6]朱风鹏,李雪,罗彦波,等.爆珠破碎对主流烟气有害成分释放量和滤嘴截留的影响[J].烟草科技,2017,50(4):37-41.Zhu Fengpeng,Li Xue,Luo Yanbo,et al.Effects of capsule crushing on cigarette filter’s filtration and delivery of seven harmful components in mainstream cigarette smoke[J].Tobacco Science&Technology,2017,50(4):37-41.
[7]张晋,张琳.一种爆珠滤棒中爆珠的微波测定方法:CN105866140A[P].2016-08-17.Zhang Jin,Zhang Lin.A microwave method for capsule in capsule filter:CN105866140A[P].2016-08-17.
[8]瞿海斌,刘晓宣,程翼宇.中药材三七提取液近红外光谱的支持向量机回归校正方法[J].高等学校化学学报,2004,25(1):39-43.Qu Haibin,Liu Xiaoxuan,Cheng Yiyu.A method for analysis of extracting solution of Panax Notoginseng using near-infrared spectroscopy and support vector regression[J].Chemical Journal of Chinese Universities,2004,25(1):39-43.
[9]王国祥,王海燕,王虎,等.基于Raman光谱和支持向量机回归的古井贡酒年份鉴别方法[J].光谱学与光谱分析,2016,36(3):729-735.Wang Guoxiang,Wang Haiyan,Wang Hu,et al.Study on recognition of liquor age of Gujing based on Raman spectra and support vector regression[J].Spectroscopy and Spectral Analysis,2016,36(3):729-735.
[10]熊宇虹,荣祺,舒明磊,等.改进的基于线性支持向量机回归的近红外光谱分析方法[J].上海应用技术学院学报(自然科学版),2016,16(3):275-279..Xiong Yuhong,Rong Qi,Shu Minglei,et al.Improved near infrared spectral analysis method based on linear support vector machine regression[J].Journal of Shanghai Institute of Technology(Natural Science),2016,16(3):275-279.
[11]黎洪利,赵朋贤,杨俊,等.紫外指纹图谱在烟用香精质量控制中的应用[J].中国烟草学报,2009,15(4):19-21.Li Hongli,Zhao Pengxian,Yang Jun,et al.Application of ultraviolet fingerprint spectrum in tobacco flavor quality control[J].Acta Tabacaria Sinica,2009,15(4):19-21.
[12]孙志涛.紫外-可见分光光度法在烟用香精香料质量评价中的应用[J].安徽农学通报,2010,16(2):144-147.Sun Zhitao.Application of ultraviolet-visible spectrum piin tobacco flavor quality evaluation[J].Anhui Agricultural Science Bulletin,2010,16(2):144-147.
[13]余志雄,周创兵,李俊平,等.基于ν-SVR算法的边坡稳定性预测[J].岩石力学与工程学报,2005,24(14):2468-2475.Yu Zhixiong,Zhou Chuangbing,Li Junping,et al.Predicting slope stability based onν-SVR algorithm[J].Chinese Journal of Rock Mechanics and Engineering,2005,24(14):2468-2475.
[14]Chang Chih-Chung,Lin Chih-Jen.Trainingν-Support Vector Regression:Theory and Algorithms[J].Neural Computation,2002,14(8):1959-1977.
[15]Huang Bo,Cai Zhihua,Gu Qiong,et al.Using Support Vector Regression for Classification[C].International Conference on Advanced Data Mining and Applications,2008,5139:581-588.
[16]Chen Muyen,Chen Chiachen,Chang Yafen.Comparison of support vector machine and support vector regression:An application to predict financial distress and bankruptcy[C].International Conference on Service Systems&Service Management,2010,53(5):1-6.
[17]吴元清,杜树新,严赟,等.水体有机污染物浓度检测中的紫外光谱分析方法[J].光谱学与光谱分析,2011,31(1):233-237.Wu Yuanqing,Du Shuxin,Yan Yun,et al.Ultraviolet spectrum analysis methods for detecting the concentration of organic pollutants in water[J].Spectroscopy and Spectral Analysis,2011,31(1):233-237.