应用BP神经网络改进熏煮香肠质构的感官评定预测  

Improvement of Prediction of Sensory Texture Evaluation of Smoked-and-cooked Sausage by BP Neural Network

作　　者：董庆利[1] 罗欣[2] 李保国[1] 李红梅[1] 李代禧[1] 

机构地区：[1]上海理工大学食品与生物技术研究所,上海200093 [2]山东农业大学食品学院,山东泰安271018

出　　处：《食品科学》2008年第12期74-78,共5页Food Science

基　　金：国家自然科学基金项目(30800864);上海高校选拔培养优秀青年教师科研专项基金项目(slg-07049)

摘　　要：应用BP神经网络对熏煮香肠质构的感官评定预测进行了改进。数学检验结果表明,建立的BP神经网络模型平方根误差(RMSE)和标准预测误差(%SEP)较低,显著低于多元回归模型,而偏差因子(Bf)和准确性因子(Af)都在可接受范围。BP神经网络可以作为较好的预测模型用于实际肉类工业中肉制品的质构感官评价,实现机械测定全部或部分代替感官评定的快速性、实时性、便捷性检测。Prediction of sensory texture evaluation of smoked-and-cooked sausage was obtained and improved with back propagation neural network （BPNN）, which had been analyzed on the basis of multiple regressions by Dong and Luo previously. It was found that the accuracy and goodness-of-fit of BPNN are higher significantly than those of multiple regressions with lower root-mean-squares error （RMSE） and standard error of prediction （SEP）, and accuracy factor （Af） and bias factor （Bf） in acceptable range. Therefore, BPNN provides a useful and accurate method for predicting sensory texture evaluation in meat industry, and is meaningful for the fast, on-time and convenient detection of instrumental texture measurement instead of sensory evaluation.

关 键 词：BP神经网络 熏煮香肠 质构 预测 

分 类 号：TS251.65[轻工技术与工程—农产品加工及贮藏工程]