Prediction of Sugar Content and Acidity of Fresh Jujubes Using Wavelet Transform and Artificial Neural Network Model
Totally 360 samples from 6 kinds of fresh jujubes (60 samples of each kind) are selected for spectral analysis. Firstly,First Derivative (FD) method and Multiple Scattering Correction (MSC) method are used for pretreatment data. Secondly,wavelet transform is applied to compress the data. Finally,fresh jujubes prediction models of sugar content and acidity are established by BP Artificial Neural Network (BP-ANN) and RBF Artificial Neural Network (RBF-ANN) respectively. The performance of the model is evaluated by the correlation coefficient and Root Mean Square Error of Prediction (RMSEP). Experimental results show that the sugar content result calculated by correlation coefficient is equal to 0.9941 and calculated by RMSEP is 1.6205. In addition,the acidity result calculated by correlation coefficient is 0.9815 and calculated by RMSEP is equal to 0.2533. Compared with Partial Least Square (PLS),BP-ANN and RBF-ANN,we find that RBF-ANN model is better than BPANN model and PLS model in the prediction of sugar content and acidity of fresh jujubes.
fresh jujube wavelet transformation artificial neural network
Jiahu Peng Xiyuan Wang Dan Liu
School of Physical and Electrical Information Engineering,Ningxia University,Yinchuan 750021,China
国际会议
西安
英文
1105-1109
2011-12-23（万方平台首次上网日期，不代表论文的发表时间）