The study proposes an one-dimensional shallow convolutional neural network combined with elastic network(1D-SCNN-EN)to predict the blood glucose concentration in the blood Raman spectrum,and obtains 106 different blood glucose spectra by Fourier transform(FT)Raman spectra.Then the study proposes an one-dimensional shallow convolutional neural network with elastic net to capture multiple deep features,and reduce the complexity of the model.1D-SCNN-EN model has better performance than traditional methods(partial least square method and support vector machine).The root-mean-square error(RMSEC)of correction set,root-mean-square error(RMSEP)of prediction set,determination coefficient(R2P)and relative analysis error(RPD)are 0.10262,0.11210,0.99403 and 12.94601,respectively.The experimental results show that compared with other regression models,the model has higher prediction accuracy and stronger robustness.In the case of a small amount of data,the model is expected to predict the blood glucose concentration through Raman spectroscopy.
关键词
血糖/拉曼光谱/浅层卷积神经网络/弹性网
Key words
Blood glucose/Raman spectrum/Shallow convolutional neural network/Elastic net
维普
万方数据