基于模糊集的玉米饮料滋味自动鉴评方法  

Automation Evaluation of Corn Juices Taste Based on Fuzzy Information

作　　者：刘晶晶[1] 孙永海[1] 陈莉[1] 孙钟雷[2] 王笑丹[1] 付天宇[1] 

机构地区：[1]吉林大学生物与农业工程学院,长春130025 [2]长江师范学院生命科学与技术学院,重庆408100

出　　处：《农业机械学报》2013年第3期147-152,共6页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金资助项目(31271861);国家高技术研究发展计划(863计划)资助项目(2008AA100802)

摘　　要：利用传感器阵列对玉米饮料滋味在模糊信息层面上实现自动化鉴评。针对玉米饮料滋味感官鉴评的不同剖面引入权值概念,利用云模型实现定性定量信息的转换和完成综合云模型基于权值差异的调整。分析传感器对玉米饮料滋味感官鉴评甜味、酸甜味、入口风味3个剖面的敏感度差异。在构建的模糊神经网络中,将针对特定剖面敏感的传感器阵列采集信息作为输入,感官鉴评云模型转化的信息作为输出,训练模糊神经网络,以得到的模糊化层中心值、模糊化层节点宽度值和模糊决策层调节参数来确定网络结构。预测分析结果表明,该系统在玉米饮料滋味模糊信息的鉴评过程中,误差率在0.002 43～0.091 77之间,效果良好。Fuzzy information of corn juices was automatically evaluated based on a sensor array. The concept of weights was introduced for different taste sensory evaluation aspects of corn juices. The conversion of the qualitative and quantitative information was achieved. At the same time, adjustment of a comprehensive cloud model was completed based on the difference of weight. According to the requirements of the different evaluation aspects of corn juices including sweetness, soursweet and flavor, sensor array were analyzed and combination of different sensor array signals were obtained. Fuzzy neural networks were built for prediction of corn juices taste fuzzy information. The information for different aspects collected from sensor array was input. The information from cloud model according to sensory evaluation was output. With training fuzzy neural network, fuzzy layer center value, the fuzzification layer node width values and fuzzy decision-making regulation parameters were obtained to determine the network structure. The forecast analysis showed that the system allowed good effect with 0. 002 43 - 0. 091 77 error rate in the process of automation evaluation of fuzzy information for corn juices.

关 键 词：玉米饮料滋味鉴评模糊神经网络 云模型 

分 类 号：TS207.3[轻工技术与工程—食品科学]