Pcomparison Of The Performance Of Five-dimensional Calibration Algorithms And The Application In Pharmaceutical Analysis

With the rapid development of modern science and technology,a large number of analytical instruments of various types continue to emerge,especially the emergence of combined instruments that can produce high-dimensional data,which brings new opportunities and challenges for analytical chemistry.Chemometrics,as a newly developed interdisciplinary subject combining traditional subjects such as mathematics,statistics and computer,has the characteristics of processing complex and high-dimensional data and being able to extract a large amount of effective information from it.In Chapter 2,we compare and evaluate the performance of three five-linear calibration algorithms which are five-way parallel factor analysis（five-PARAFAC）,alternating fitting weighted residue quinquelinear decomposition（AFWRQQLD）and alternating quinquelinear decomposition（AQQLD）.We use a five-dimensional HPLC-DAD simulation data and a five-dimensional fluorescence spectrum real data to evaluate the analytical performance of these three algorithms in terms of convergence performance,collinearity effect,and number of components.Among them,Five-PARAFAC which in the form of one-way fully expanded matrix has slow convergence speed,prone to recessionary solutions,and is more sensitive to the group fraction required by the fitting model.AQQLD which based on the slicing matrix is weak to noise tolerance.AFWRQQLD,which is based on two different unidirectional and incompletely expanded forms,has better analytical capabilities than the former two.This conclusion provides a reference for the development of new five-dimensional correction algorithms.In Chapter 3,flavonoids,as multifunctional compounds with medicinal value,play an important role in the treatment of human diseases.However,its fluorescence signal is easily affected by factors such as pH value and solution background interference.The traditional fluorescence analysis methods cannot achieve non-interference quantitative detection in the human environment,so it is urgent to achieve a new fluorescence detection method to achieve quantitative detection.In this chapter,we use five-dimensional algorithm AFWRQQLD and five-dimensional emissions-excitation-pH-human serum volume fluorescence array,directly enable to analysis of 7-methoxyflavonoids in complex human serum samples merely by sample pretreatment.The recovery rate of 7-methoxyflavonoids in human serum was 105.9±6.6%,and the detection limit was 0.002μg m L^-1.At the same time,we combine with other quality factors to verify the performance of the method.The experimental results show that this method can effectively avoid the influence of interference signals on the determination results in complex environment,and provides an effective way for the monitoring of human serum drug concentration.In Chapter 4,7-Hydroxyflavanone is widely present in plants and has become an important research object with pharmacological activity in natural organics.Therefore,it is necessary to study its detection in human environment.Aiming at the problems of strong fluorescence spectrum overlap between 7-hydroxyflavanone and the complex background matrix of human serum which has the background matrix effect and leads to the enhancement of the fluorescence signal of the target analyte,we use the AFWRQQLD five-dimensional correction algorithm to analyze the emission-excitation-pH-human serum volumetric five-dimensional fluorescence matrix.With the high-order advantages of the five-dimensional calibration algorithm,the quantitative determination of 7-hydroxyflavanone in human serum can be achieved by simply preprocessing the experimental samples.The average recovery rate of 7-hydroxyflavanone in human serum is 99.7～108.2%.Combining with the calculation results of quality factor,the algorithm has a good ability to eliminate the background interference and matrix effect in human serum,and the detection results are accurate and reliable.The quantitative determination method is simple and accurate,and provides a reliable quantitative method for the detection of drug molecules in human serum.