摘要
采用近红外光谱对大环内酯类抗生素罗红霉素制剂进行分析,然后对全光谱波长使用后向区间偏最小二乘法进行选择预处理,接着将其压缩为主要成分,运用BP人工神经网络进行预测模型的构建,并对罗红霉素药物的主要成分进行分析。该模型的建立能够对噪声等偶发因素造成的瞬时扰动进行有效抵御,同时还可减少建模使用的波长数、运算时间,提高预测精度。
The macrolide antibiotic roxithromycin by near infrared spectroscopy, then, the back propagation partial least squares（PLS） method is used to preprocess the full spectral wavelength, then the compression is the main component, forecast model of BP artificial neural network method, and the main ingredient of roxithromycin drug analysis. The model can effectively resist the instantaneous disturbance caused by noise and other accidental factors, and can also reduce the number of wavelengths used, the operation time and the prediction accuracy.
出处
《中国继续医学教育》
2017年第16期190-192,共3页
China Continuing Medical Education
