1D convolutional neural network for the discrimination of aristolochic acids and their analogues based on near-infrared spectroscopy

摘要: Chinese herbs containing aristolochic acids (AAs) have been implicated in renal failure and urothelial carcinoma. The detection of AAs their analogues is significant for the correct use drugs clinical medicine. Traditional discrimination methods based on near-infrared spectroscopy (NIRS) technique generally employ wavelength selection algorithms to eliminate redundant wavelengths before constructing shallow learning classifier. However, are defective increasing complexity model depend skilled expertise knowledge. To avoid these drawbacks, an end-to-end 1-dimensional convolutional neural network (1D-CNN) basis NIRS developed distinguish this paper, which extracts feature from original input data instead using manually. Moreover, back propagation artificial network, support vector machine, principal component analysis combined with machine t-distributed stochastic neighbor embedding established make comparisons proposed model, respectively. T-distributed (t-SNE) visualization results indicate that 1D-CNN has excellent ability. experimental comparison show generalized performance outperforms traditional classifier or algorithms. Thus, designed easy effective qualitative tool analogues.