Research On Quality Classification Of Auricularia Auricula Polysaccharide Based On Support Vector Machine Optimized By Genetic Algorithm

Auricularia auricula is a common edible fungus in China.It is rich in various nutrients and is loved by the masses.In recent years,along with the deepening of the research on the composition of Auricularia auricular,the research of Auricularia auricular polysaccharide(APP)has been highly valued by researchers.As a high-purity polysaccharide extract from Auricularia auricula,Auricularia auricula polysaccharide has the functions of anti-aging,anti-carcinogenic factors(as well as tumors),anti-coagulation,reducing blood lipids and blood sugar.Auricularia auricula polysaccharide has a deep potential for medical development.Traditional methods for the detection and classification of polysaccharides from Auricularia auricula use chemical extraction methods.However,chemical reagents always corrode samples,which cannot accurately reflect the composition information.Moreover,the operation is complicated and the detection period is long.In order to adapt to high-efficiency industrial production,a kind of fast,accurate and safe quality classification technology of auricularia auricula polysaccharides is needed.As an analysis and detection technology with little pre-treatment,little waiting,no pollution,Near infrared spectra analysis is being used more and more in the detection of agricultural products and food quality.For the near-infrared spectra of different types of Auricularia auricular,the absorption frequencies at the same waveform frequency are not the same.By comparing the wave values at the same frequency of the near-infrared spectrum,the content of the Auricularia auricular polysaccharide in the corresponding sample can be known.The purpose of non-destructive testing.As an excellent simulation and modeling method,support Vector Machine(SVM)has the advantages of simple structure and strong generalization ability.It has been applied to character recognition,image classification,biochemical composition analysis and so on.However,the key parameters of SVM have a great impact on classification and recognition.Because the research on parameter selection is not yet mature,the accuracy of model recognition and classification is often low.In order to solve the above problems,this paper uses genetic algorithm to optimize parameters and establish a support vector machine model.Firstly,non-destructive detection of polysaccharide components in different seasons and different cultivation methods of the Auricularia auricula samples was carried out.The standard data was used to pre-process the spectral data by standard normal variable transformation(SNV)+SG smoothing to eliminate the effects of spectral drift,surface scattering and noise;Then,the feature wavelength is extracted by successive projections algorithm(SPA),which can reduce the influence of uncorrelated variables on the classification effect,and extract the spectral features to construct the support vector machine model.Finally,the genetic algorithm(GA)is used to optimize the kernel function parameters.The experimental results show that the classification accuracy of the SVM model based on GA optimization is 97.5%,the accuracy rate is 95.6%,the recall rate is 100%,the F1-score is 0.974,and the AUC is 0.941.The experimental results are in line with the expected goals,indicating that the GA-SVM model can be used to classify and measure the polysaccharide content of Auricularia auricula in practical applications.