浏览全部资源
扫码关注微信
江苏大学 食品与生物工程学院,江苏 镇江 212013
郑开逸,博士,副研究员,研究方向:食品无损检测,E-mail:kaiyizhengjsu@126. com
收稿日期:2024-08-02,
修回日期:2024-09-05,
录用日期:2024-09-11,
纸质出版日期:2025-04-15
移动端阅览
赵丽娜,沈烨,商显文,陈智扬,石吉勇,李志华,黄晓玮,郑开逸.基于高光谱结合半监督回归的肴肉硫代巴比妥酸反应物的测定[J].分析测试学报,2025,44(04):708-713.
ZHAO Li-na,SHEN Ye,SHANG Xian-wen,CHEN Zhi-yang,SHI Ji-yong,LI Zhi-hua,HUANG Xiao-wei,ZHENG Kai-yi.Detection of TBARS in Yao-meat Based on Hyperspectral Imaging Combined with Semi-supervised Regression[J].Journal of Instrumental Analysis,2025,44(04):708-713.
赵丽娜,沈烨,商显文,陈智扬,石吉勇,李志华,黄晓玮,郑开逸.基于高光谱结合半监督回归的肴肉硫代巴比妥酸反应物的测定[J].分析测试学报,2025,44(04):708-713. DOI: 10.12452/j.fxcsxb.240802273.
ZHAO Li-na,SHEN Ye,SHANG Xian-wen,CHEN Zhi-yang,SHI Ji-yong,LI Zhi-hua,HUANG Xiao-wei,ZHENG Kai-yi.Detection of TBARS in Yao-meat Based on Hyperspectral Imaging Combined with Semi-supervised Regression[J].Journal of Instrumental Analysis,2025,44(04):708-713. DOI: 10.12452/j.fxcsxb.240802273.
该文以肴肉的硫代巴比妥酸反应物(TBARS)为新鲜度指标,通过高光谱结合半监督学习进行预测。在数据集中,高光谱数据为
X
,TBARS含量数据为
y
值。同时,将整个样本集合分为校正集、验证集、独立测试集。其中,校正集用于建立模型,以预测验证集和独立测试集。在校正集中,既有
X
,又有
y
的样本称为有标样本;而仅有
X
,没有
y
的样本称为无标样本。验证集和独立测试集中的每一个样本均为有标样本。验证集仅用于调节校正集建立模型的参数,不参与建模。独立测试集则不参与建模也不参与调节参数,仅用于测试模型最终的结果。文中校正集样本数为233,其中有标样本48个,无标样本185个;验证集和独立测试集样本数均为12。在建模过程中,先用校正集中的有标样本建立
X
和
y
的模型;然后用此模型预测无标样本,预估其
y
值。此时,校正集中所有样本均为有标样本。最后,基于校正集中的所有样本建模,构建模
型用于预测。所构建的两种模型的参数存在差异,均通过验证集进行优化。结果显示:支持向量机回归(SVR)的建模效果较好,同时,SVR算法结合半监督学习可以获得较高的预测精度。在无标样本的选择中,相比基于全部无标样本的方法,基于距离法选择的无标样本可以获得更低的预测误差。
Yao-meat is the traditional food of Zhenjiang. However,the Yao-meat is easy to spoil,thus,it is necessary to monitor its freshness. The traditional monitoring methods including physical and chemical tests have the disadvantages,such as complex operation and time-consuming preparation. Therefore,this paper used hyperspectral imaging as a non-destructive detection method to measure the freshness of Yao-meat. The thiobarbituric acid reactive substance(TBARS) was set as the indicator of freshness,and predicted by the hyperspectral imaging combined with semi-supervised regression. For the dataset,the hyperspectral images and the values of TBARS were assigned as
X
and
y
. Meanwhile,the dataset was divided as three parts,including calibration,validation and independent prediction sets. Among them,the calibration set was used to build model and predict samples in validation and prediction sets. In calibration set,the samples with and without
y
values were called as labeled and unlabeled ones. The samples in validation and independent prediction sets were all labeled ones. The validation set was applied to adjust the parameters of model rather than calibration. Meanwhile,the independent test set was used to test the final results of model rather than calibration and parameters adjustment. In this paper,the sizes of calibration,validation and independent prediction sets were 223,12 and 12,respectively. Among the calibration set,the numbers of labeled and unlabeled samples were 48 and 185,respectively. During calibration,the labeled samples with
X
and
y
values were computed to generate model initially. Then the unlabeled samples in calibration set were predicted by the mode
l. At this time,all the samples in calibration set were labeled. Finally,the whole samples in calibration set were applied to build a new model for prediction. The parameters of the two models may be different to each other and were both optimized by validation sets. In this paper,the nonlinear models were executed for calibration,including support vector regression(SVR),Gaussian process regression(GPR),artificial neural network(ANN) and extreme learning machine(ELM). The results show that compared with different calibration models,the SVR can achieve good results. In the same time,SVR combined with semi-supervised regression can improve prediction accuracy. Moreover,selecting key unlabeled samples based on distances can reduce prediction errors.
Nie X R . Cook. Knowl . (聂晓瑞. 烹调知识) , 2007 ,( 6 ): 35 - 36 .
Wang G Y . Sichuan Cuisine (王广宇 . 四川烹饪) , 2002 ,( 2 ): 19 .
Li D L , Li H B , Liu E Z , Ren F Z , Pang R P . Meat Ind. (李大龙 , 李海宾 , 刘尔卓 , 任发政 , 庞瑞鹏 . 肉类工业) , 2015 ,( 1 ): 21 - 23 .
Waimin J , Gopalakrishnan S , Heredia-Rivera U , Kerr N A , Nejati S , Gallina N L F , Bhunia A K , Rahimi R . ACS Appl. Mater. Interfaces , 2022 , 14 ( 40 ): 45752 - 45764 .
Franco M R , da Cunha L R , Bianchi R F . Sens . Actuators B , 2021 , 333 : 129466 - 129471 .
Katiyo W , de Kock H L , Coorey R , Buys E M . LWT , 2020 , 128 : 109468 - 109476 .
Dabadé D S , Yessoufou N , Adido L , Azokpota P , Hounhouigan D J . Int. J. Food Microbiol. , 2023 , 405 : 109468 - 109476 .
Pereira P F M , de Sousa Picciani P H , Calado V , Tonon R V . Food Packag . Shelf , 2023 , 36 : 101049 - 101058 .
Chen Y , Ma F , Wu Y , Tan S , Niu A , Qiu W , Wang G . Food Microbiol. , 2023 , 115 : 104311 - 104318 .
Bassey A P , Chen Y , Zhu Z , Odeyemi O A , Gao T , Olusola O O , Ye K , Li C , Zhou G . Food Control , 2021 , 130 : 108383 - 108396 .
Liu H , Ji Z , Liu X , Shi C , Yang X . Food Chem. , 2020 , 321 : 126628 - 126633 .
Beya M M , Netzel M E , Sultanbawa Y , Smyth H , Hoffman L C . Meat Sci. , 2023 , 204 : 109268 - 109277 .
Zhang L , Wang Y , Wei Y , An D . Food Chem. , 2022 , 370 : 131047 - 131055 .
Peng W , Beggio G , Pivato A , Zhang H , Lü F , He P . Renew. Sust. Energ. Rev. , 2022 , 165 : 112608 - 112617 .
Khamsopha D , Woranitta S , Teerachaichayut S . Food Control , 2021 , 123 : 107781 - 107788 .
Liu L , Zareef M , Wang Z , Li H , Chen Q , Ouyang Q . Food Chem. , 2023 , 412 : 135505 - 135512 .
Ouyang Q , Rong Y , Wu J , Wang Z , Lin H , Chen Q . Food Chem. , 2023 , 420 : 136078 - 136086 .
Sanchez P , Arogancia H , Boyles K , Pontillo A , Ali M . Appl. Food Res. , 2022 , 2 ( 2 ): 100147 - 110163 .
Feng C , Makino Y , Oshita S , Garcia M , Juan F . Food Control , 2018 , 84 : 165 - 176 .
Fu X , Chen J . Compr. Rev. Food Sci. Food Saf. , 2019 , 18 ( 2 ): 535 - 547 .
Antequera T , Caballero D , Grassi S , Uttaro B , Perez-Palacios T . Meat Sci. , 2021 , 172 : 108340 - 108351 .
Fan N , Ma X , Liu G , Ban J , Yuan R , Sun Y . J. Food Compos. Anal. , 2021 , 103 : 104110 - 104119 .
Cheng J , Sun J , Yao K , Dai C . Food Control , 2023 , 153 : 109940 - 109948 .
Ramesh S , Srivastav V , Alapatt D , Yu T , Murali A , Sestini L , Nwoye C I , Hamoud I , Sharma S , Fleurentin A , Exarchakis G , Karargyris A , Padoy N . Med. Image Anal. , 2023 , 88 : 102844 - 102862 .
Barbero-Aparicio J A , Olivares-Gil A , Rodríguez J J , García-Osorio C , Díez-Pastor J F . Inform . Fusion , 2024 , 102 : 102035 - 102044 .
Gomes H M , Grzenda M , Mello R , Read J , Le Nguyen M H , Bifet A . ACM Comput. Surv. , 2022 , 55 ( 4 ): 1 - 22 .
Liu K , Liu H , Wang T , Hu G , Ward T E , Chen C L P . IEEE T Neur. Net. Lear. , 2023 , 35 ( 9 ): 1 - 12 .
Kutsanedzie F Y H , Agyekum A A , Annavaram V , Chen Q . Food Chem. , 2020 , 315 : 126231 - 126240 .
Peng H G , Jin Y , Zhan Y G , Chen Y Q , Feng X B , Qian F C , Huang G , Huang T J , Li J . J. Instrum. Anal. (彭海根,金楹,詹莜国,陈雅琼,封幸兵,钱发聪,黄果,黄天杰,李杰. 分析测试学报), 2020 , 39 ( 10 ): 1305 - 1310 .
Sun P , Xiang C Q , Chen Q W , Jia B , Li X Y , Chen W X , Qiao W L , Xiao X . J. Instrum. Anal. (孙鹏,向超群,陈启文,贾彬,李欣怡,陈炜璇,乔卫林,肖雪. 分析测试学报), 2024 , 43 ( 4 ): 637 - 642 .
0
浏览量
36
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构