Detection Of Environmental Samples By Infrared Model Combined With Chemometrics

With the development of modern analytical and testing instruments,people can obtain a large amount of data in a short time,but how to effectively analyze these data becomes more difficult.Chemometrics can reveal chemical information or rules hidden in chemical measurement data,and it is an important tool for chemical workers to analyze multidimensional data.At present,chemometrics is widely used in environmental chemistry,medical chemistry,food chemistry and other fields.This paper summarizes the commonly used chemometrics methods,including support vector machine,artificial neural network,K-nearest neighbor method,soft independent modeling of class analogy,Fisher linear discriminant method,cluster analysis,principal component analysis and partial least squares method.The common preprocessing methods of spectral data are introduced,including baseline correction,smoothing,mean centralization,standardization,derivative,wavelet transform,multivariate scattering correction and so on.Qualitative and quantitative analysis of compound paracetamol and amantadine hydrochloride and adulterated oil was carried out by chemometrics.The specific research contents and results were as follows:(1)Ultraviolet spectrophotometry combined with chemometrics was used to analyze the ultraviolet spectra of compound paracetamol and amantadine hydrochloride.The compound paracetamol and amantadine hydrochloride of three manufacturers was classified by uncorrelated linear discrimination analysis method,and the classification result was good.The contents of paracetamol and chlorphenamine maleate in 6 different batches of compound paracetamol and amantadine were predicted by partial least squares method.The average recoveries of paracetamol and chlorphenamine maleate were 93.0%～100.7% and 93.54%～98.92%respectively.The results were accurate.(2)Fourier transform infrared spectroscopy and chemometrics were used to qualitatively analyze the infrared spectra of different quality oils.The three kinds of oils were classified by principal component analysis and uncorrelated linear discrimination analysis,and the classification result was good.The cluster analysis method was used to qualitatively analyze the oils.The correct classification rates of oils obtained by euclidean distance method,urban block distance method and chebyshev distance method were 98%,96% and 94%,respectively.The classification result was good.The chemometrics provides a more effective technical means for the qualitative andquantitative detection of complex environmental samples.With the development of computer technology,the chemometrics will certainly be more widely used,and its prospects will be even broader.