基于PCA-GM-BP神经网络的猪肉价格预测分析

针对猪肉价格上下波动呈非线性关系和影响因素复杂等难以预测的问题,提出了基于PCA-GM-BP神经网络预测模型对猪肉价格进行有效预测.以2010年1月-2018年12月的月度价格数据作为样本,共计108组数据,利用PCA对影响猪肉价格变化的12种因素进行降维处理,选用对猪肉价格的主要累积贡献率超过96%的5个主成分,构建PCA-GM-BP神经网络猪肉价格预测模型.结果表明:与传统的BP神经网络、GM-BP神经网络预测模型相比,PCA-GM-BP神经网络预测模型在提高聚类效果的同时,增加了预测结果的精确性,对我国猪肉价格预测具有更高的适用性与参考价值性.     

Pork fat peroxidation by gamma-irradiation

In this paper, pork fat peroxidation by γ-irradiation and the possible effects of oxygen, UV-irradiation and storage after the γ-irradiation have been investigated. It has been found that the level of peroxides in irradiated pork increases linearly with the increasing absorbed dose. The chemical yield of peroxides formed in the irradiated fat is about 4.2 and independent on the sample temperature or absorbed dose rate, but dependent on storage time of sample before γ-irradiation. The irradiated pork exhibits some unusual features as following: 1) the peroxide content in irradiated pork is higher than that in unirradiated one; 2) the peroxide content in irradiated pork increases gradually on storage and is essentially constant in unirradiated one, which is very useful for the detection of irradiated pork; 3) the further peroxidation in irradiated pork is much more susceptible to UV radiation than that in unirradiated pork.     

应用近红外漫反射光谱对猪肉肉糜进行定性定量检测研究

利用傅里叶变换近红外漫反射光谱结合不同数学建模算法对不同部位取样的猪肉肉糜进行定性判别建模,并建立猪肉肉糜品质指标脂肪、蛋白质和水分含量的定量检测模型。结果表明:不同部位取样猪肉肉糜判别分析PLSDA模型性能良好,最优模型校正集判别正确率为100%,预测集判别正确率为96%;比较两种方法结合,不同光谱预处理建立各品质指标的定量模型,LS-SVM模型性能优于PLSR模型,脂肪和水分含量最佳预测模型校正及预测相关系数r均高于0.9,蛋白质含量最优模型校正及预测相关系数r,RMSEC,RMSEP和RMSECV分别为0.722,0.593,1.595,1.550和1.888,模型精度需进一步提高。研究表明利用傅里叶变换近红外漫反射光谱快速判别不同部位猪肉肉糜的方法是可行的,脂肪和水分含量定量分析模型从预测精度、稳定性及适应性考虑均具一定的通用性,具有良好的市场应用前景。     

多壁碳纳米管固相萃取净化-高效液相色谱法测定猪肉和鸡肉中的磺胺多残留

建立了多壁碳纳米管为吸附剂的固相萃取净化-高效液相色谱-紫外检测测定猪肉和鸡肉中多种磺胺类药物多残留的方法。样品采用乙腈提取,多壁碳纳米管固相萃取净化,NaH₂PO₄缓冲溶液（pH 5.5~6.0）溶解上样,5%（v/v）丙酮-正己烷淋洗,丙酮-二氯甲烷（1：1,v/v）洗脱。色谱分离以50 mmol/L NaH₂PO₄-乙腈（7：3,v/v）为流动相,方法的线性范围为0.01~1.00 mg/L,线性相关系数大于0.998,检出限（LOD）为0.003 mg/L,定量限（LOQ）为0.01 mg/L。在0.02~0.2 mg/kg添加范围内,9种磺胺类药物的回收率高于70%,RSD低于8%,表明多壁碳纳米管对磺胺类药物具有较强的吸附富集能力。该方法简便、准确可用于动物组织及产品中磺胺药物残留的检测。     

高效液相色谱法快速检测猪组织中泰妙菌素残留

应用高效液相色谱法测定猪组织中泰妙菌素残留.试样经磷酸盐缓冲液(pH 8.0)-乙酸乙酯混合提取,离心后加0.1%酒石酸,40 ℃旋转蒸发,除尽有机相,加正已烷除脂,HLB柱净化后,氮气吹干,0.0125%甲酸溶解后HPLC检测.流动相为乙腈-0.0125%甲酸(42∶ 58,V/V).紫外检测波长为210 nm.对猪肉(猪肝)进行15(25)～500 μg/kg的药物添加回收实验,结果表明,泰妙菌素的平均回收率为80.1%～92.7%; 日内相对标准偏差为2.0%～10.2%;日间相对标准偏差为5.9%～13.9%.按信噪比S/N=3计算,猪肉和猪肝中泰妙菌素的检出限(LOD)分别为4.5和7.5 μg/kg.     

An Improved Residual Network for Pork Freshness Detection Using Near-Infrared Spectroscopy

Effective and rapid assessment of pork freshness is significant for monitoring pork quality. However, a traditional sensory evaluation method is subjective and physicochemical analysis is time-consuming. In this study, the near-infrared spectroscopy (NIRS) technique, a fast and non-destructive analysis method, is employed to determine pork freshness. Considering that commonly used statistical modeling methods require preprocessing data for satisfactory performance, this paper presents a one-dimensional squeeze-and-excitation residual network (1D-SE-ResNet) to construct the complex relationship between pork freshness and NIRS. The developed model enhances the one-dimensional residual network (1D-ResNet) with squeeze-and-excitation (SE) blocks. As a deep learning model, the proposed method is capable of extracting features from the input spectra automatically and can be used as an end-to-end model to simplify the modeling process. A comparison between the proposed method and five popular classification models indicates that the 1D-SE-ResNet achieves the best performance, with a classification accuracy of 93.72%. The research demonstrates that the NIRS analysis technique based on deep learning provides a promising tool for pork freshness detection and therefore is helpful for ensuring food safety.     

基于光谱技术和支持向量机的生鲜猪肉水分含量快速无损检测

为实现生鲜肉水分含量的快速无损检测,在波长350～1 700 nm范围内采集生鲜猪肉98个样本的可见近红外反射光谱。经中值平滑滤波、多元散射校正和一阶微分复合预处理方法对原始光谱进行降噪处理。将样本数据随机分为训练集和测试集,以训练集交叉验证网格搜索法确定最佳惩罚参数,利用径向基核函数的支持向量机算法建立了支持向量机预测模型,并与偏最小二乘回归建模法进行比较。用径向基核函数的支持向量机算法所建模型对生鲜肉水分含量进行预测的结果为：训练集的预测相关系数R_c为0.96、标准差SEC为0.32,测试集的预测相关系数R_v为0.87、标准差SEV为0.67。实验结果证实用支持向量机所建模型适合于生鲜猪肉水分含量的无损快速检测。     

优质猪肉供应链中养猪场选择评价指标体系研究

通过科学评价选择养猪场是建设和管理优质猪肉供应链倒丶?本文在评述国内外关于供应链合作伙伴评价指标体系构建方法的基础上,提出了“理论分析、专家咨询;问卷调查、统计分析;融合调整,逐步完善“的评价指标体系设计思路,构建出了一套适合于优质猪肉供应链核心企业选择养猪场的评价指标体系,包括地理环境、质量保障、生产能力和合作程度等4类12个指标.     

基于脂肪酸差异的肉制品中猪源性成分鉴别方法的研究

建立了肉制品中脂肪酸的分析检测技术,通过分析脂肪酸含量、种类以及片段特征指纹图谱,实现了基于脂肪酸差异的肉制品中猪源性成分的鉴别。采用气相色谱法测定了大量肉制品,利用保留时间和质谱鉴定了36种脂肪酸成分,选取较为稳定的18种脂肪酸数据进行统计分析,确定不同油脂脂肪酸的差异,结合特征图谱,最终确定可用于鉴别猪源性成分的8种脂肪酸。该方法可用于确定肉制品中是否含有猪源性成分,方法简单、快速、成本低,辨别度较高,在猪源性成分鉴别中具有一定的应用前景。     

Determination of macrolide antibiotics residues in pork using molecularly imprinted dispersive solid‐phase extraction coupled with LC–MS/MS

A class‐specific macrolide molecularly imprinted polymer was synthesized by precipitation polymerization using tulathromycin as the template and methacrylic acid as the functional monomer. The polymers revealed different specific adsorption and imprinting factor for macrolides with different spatial arrangement of side chains as well as lactonic ring size. And the molecularly imprinted polymer possessed maximum adsorption capacity (54.1 mg/g) and highest imprinting factor (2.4) toward 15‐membered ring azithromycin. On the basis of molecularly imprinted polymer dispersive solid‐phase extraction, a rapid, selective, and reproducible method for simultaneous determination of seven macrolide antibiotics residues in pork was established by using liquid chromatography with tandem mass spectrometry. At spiking levels of 5, 10, 25, and 100 μg/kg, average recoveries of seven macrolides ranged from 68.6 to 95.5% with intraday and interday relative standard deviations below 8%. The limits of detection and limits of quantification were 0.2–0.5 and 0.5–2.0 μg/kg, respectively.