摘要
借助电子鼻检测存储60、120、180、240、300、360 d的黄山毛峰茶香气信息,根据电子鼻各传感器响应曲线变化特点,选取出1组能够表征不同香气信息的基本特征变量,分别采用主成分回归(PCR)、偏最小二乘回归(PLS)和BP神经网络(BPNN)方法,建立茶叶存储时间的预测模型。测试样本集对3种预测模型的检验结果表明:PCR、PLS、BPNN模型的预测标准误差分别为10.05、6.04、3.21d;最大预测相对误差分别为11.03%、7.02%、5.89%;平均预测相对误差分别为6.73%、4.74%、3.62%;预测值与实际值之间的决定系数R2分别为0.862、0.896、0.987。3种模型都能较好地对茶叶存储时间进行预测,相比较而言,BPNN模型性能最优,PLS模型性能优于PCR模型。
A recognition methods for storage time of tea was set up based on the Huangshanmaofeng tea under storage time of 60, 120, 180, 240, 300 and 360 d detected by electronic nose. According to response curves of electronic nose, a set of essential characteristic variables were selected. On the basis of these variables, principle component regression(PCR), partial least squares regression(PLS) and back propagation neural network(BPNN) was applied to build the prediction model for storage time of tea, respectively. Three prediction models were validated by test sample set. The results indicated that standard error of prediction of PCR, PLS and BPNN models were 10.05, 6.04 and 3.21 d, respectively;the maximum relative error 11.03%, 7.02% and 5.89%, respectively;the mean relative error 6.73%, 4.74%, and 3.62%, respectively;determination coefficient between predicted value and real value 0.862, 0.896 and 0.987, respectively. All of the models could predict storage time of tea well. BPNN was the model with the best performance and PLS is better than PCR.
作者
薛大为
杨春兰
XUE Dawei;YANG Chunlan(School of Electronics and Electrical Engineering,Bengbu University,Bengbu,Anhui 233030,China)
出处
《湖南农业大学学报（自然科学版）》
CAS
CSCD
北大核心
2019年第2期218-224,共7页
Journal of Hunan Agricultural University(Natural Sciences)
基金
安徽省高校自然科学研究重点项目(KJ2018A0574)
安徽省高校优秀青年骨干人才国内访学研修项目(gxfx2017133)
关键词
电子鼻
茶叶存储时间
多传感器融合
主成分回归
偏最小二乘回归
BP
神经网络
electronic nose
storage time of tea
multi-sensor fusion
principle component regression
partial least squares regression
back propagation neural network
