Study On Quality Risk Assessment Of Meat Products Cold Chain Logistics Based On BP Neural Network

With the improvement of people's living standard and logistics technology,the quality problem of meat products in cold chain logistics is widely concerned.However the development of technology and management of cold-chain logistics in China is relatively backward compared with developed countries,there are still quality risks in some links of cold chain logistics.In order to evaluate the degree of risk of the enterprise,provide suggestions for enterprise operation,and provide consumers with risk awareness,this thesis evaluates the quality risk of meat products in cold chain logistics.Firstly,the background and significance of the topic are introduced,the research status of domestic and foreign are described,the framework of thesis is listed.The concepts of cold chain logistics is explained,and quality risks of cold chain logistics is explained.The evaluation method of quality risk of meat products in cold chain logistics.is determined.Secondly,the operation process,management,technology and environment of cold chain logistics of meat products are analyzed in detail,and risk factors for each link of cold chain logistics are identified.By analyzing the importance and correlation of indicators,26 risk assessment indicators are finally determined.Then the quality risk assessment index system for meat products in cold chain logistics is established.Thirdly,according to the applicability and advantage analysis of BP neural network risk assessment model,the three-layer BP neural network evaluation model is established.In order to improve the efficiency of model training,principal component analysis is used to conduct data dimensionality reduction.When the cumulative contribution rate reaches 81%,8 principal components are proposed from the original data.Then the processed sample data are input into the BP neural network model..Finally,a portion of sample is used as training data in MATLAB for network training,and another portion for prediction data.Through the analysis on error and accuracy of the model,an accurate risk assessment model is obtained.In addition,taking enterprise A as an example for empirical research,and its data is substituted into the assessment model.Eventually,the risk level of enterprise A is obtained,and some suggestions for reducing risks are put forward.