论文部分内容阅读
农药活性成分的快速测定已经成为农药质量监控的一个大趋势。通过融合甲维盐制剂近红外和中红外得光谱数据,旨在用数据融合的方法建立一种快速可靠的测定甲维盐制剂活性成分的方法。采用了将偏最小二乘回归法与数据融合相结合,以及用竞争自适应重加权采样法来选择偏最小二乘回归中的有效变量的方法。与近红外和中红外各自建立的模型相比,数据融合在吸取了近红外光谱和中红外光谱相互补充的信息后,具有协同效应的模型效果有了很大的提高。同时,证实了竞争自适应重加权采样法在建模过程中是一个使得模型更加简单高效的有效的变量选择技术。研究结果表明在吸收了不同来源的多种信息之后的数据融合是一种能提高模型效果的很有效的建模方法。数据融合策略的可行性使得测定低浓度(0.1%~1.0%)样品能获得更好的结果,而且结合了变量筛选算法的对近红外和中红外光谱的数据融合,是一个很有前景的测定商业农药制剂中有效成分的方法。最后建立了一种基于近红外光谱和中红外光谱数据融合来测定商业甲维盐制剂的有效成分的方法。
Rapid determination of pesticide active ingredients has become a major trend of pesticide quality control. The data of near-infrared and mid-infrared spectra were obtained by means of the fusion of the carbaryl salt preparation and the aim of the data fusion method was to establish a fast and reliable method for the determination of the active ingredient of the carbaryl salt preparation. The method of combining partial least squares regression with data fusion and the selection of valid variables in partial least squares regression with competitive adaptive weighted weighted sampling method are adopted. Compared with the models established by both near-infrared and mid-infrared, the data fusion has greatly improved the model effect with the synergistic effect after absorbing the complementary information of near infrared spectrum and mid-infrared spectrum. At the same time, it is confirmed that competitive adaptive weighted sampling method is an effective variable selection technique that makes the model simpler and more efficient in the modeling process. The results show that data fusion after absorbing a variety of information from different sources is a very effective modeling method to improve the effectiveness of the model. The feasibility of a data fusion strategy makes it possible to obtain better results at low concentrations (0.1% -1.0%) of the sample, and incorporates a data fusion of the near-infrared and mid-infrared spectra of the variable screening algorithm, which is a promising assay Methods of commercial active ingredients in pesticide formulations. Finally, a method based on near infrared spectroscopy and mid-infrared spectroscopy data fusion to determine the active ingredients of commercial carbaryl formulations was established.