Detection Of Quinolone Antibiotics Based On Support Vector Machine And Fluorescence Spectroscopy

Antibiotics are commonly used in clinical medicine as well as animal husbandry due to their good antibacterial effect,but in recent years,the problem of antibiotic residues and pollution has become increasingly serious,which has attracted people’s attention.In order to reduce diseases and increase production,some farmers overuse antibiotics in the breeding process,which makes antibiotics remain in animal-derived foods.Long-term consumption of such foods will seriously harm human health.Therefore,it is of great significance to explore an efficient method to detect the content of antibiotics.This paper uses intelligent algorithm optimization support vector machine combined with fluorescence spectrometry to classify and predict the concentration of three quinolone antibiotics,realizes qualitative and quantitative analysis of quinolone antibiotics.The main research contents of this paper are as follows:(1)Starting from the background and advantages of the application of three-dimensional fluorescence spectrometry in the field of antibiotic detection,the feasibility of using three-dimensional fluorescence spectrometry for the analysis of quinolone antibiotics was discussed,taking into account the principles and conditions of fluorescence luminescence and the structural characteristics of quinolone antibiotics.(2)Use FS920 steady-state fluorescence spectrometer to obtain the fluorescence spectra of the single component and mixed samples of three quinolone antibiotics,including fluoroquine,ciprofloxacin and levofloxacin,at different concentrations,and analyze and study them to obtain the position of fluorescence characteristic peak,fluorescence intensity and other information.Comparing the mixed solution with the single substance spectrogram,it is found that the three antibiotics in the mixed solution interact with each other,and utilizing only three-dimensional fluorescence spectroscopy technology to classify and predict the concentration of antibiotics is challenging.An effective detection model is required to achieve qualitative and quantitative analysis of antibiotics.(3)Based on the classification principle of support vector machine,the parameter finding process of support vector machine is optimized by chicken swarm optimization algorithm,and the CSO-SVM(chicken swarm optimization algorithm to optimize support vector machine detection model)is established to classify and identify the mixed solution of antibiotics,and the prediction results are compared with the PSO-SVM(particle swarm optimization algorithm to optimize support vector machine detection model).The results show that the CSO-SVM detection model has better fitness curve and prediction accuracy.(4)The regression analysis principle of support vector machine was analyzed,the optimization process of support vector machine was optimized by improved chicken swarm optimization algorithm,and the ICSO-SVM(improved chicken swarm optimization algorithm to optimize support vector machine detection model)was established to predict the concentration of fluoromethylquine under the complex environment of three antibiotics mixing.The high recovery rate and low mean square error indicated that the ICSO-SVM model could quantitative analysis the antibiotics in mixed solution in the presence of interfering antibiotics and had the best performance compared with the CSO-SVM and PSO-SVM.(5)The standard addition method and ICSO-SVM model were used to predict the concentration of fluoroquine in the background of honey.The results showed that the model could realize the quantitative detection of antibiotics in the mixed system.