Research On Parameter Optimization Of Support Vector Machine Based On Improved Fruit Fly Optimization Algorithm

With the development of Internet technology,countless images are produced in daily life,and how to effectively classify images is a research hotspot.Support vector machine(SVM)is based on statistical learning theory,which has its unique advantages in small sample classification and has been applied in many fields.In the application of SVM,the selection of parameters has an extremely important influence on the classification accuracy and generalization ability of SVM.Therefore,the parameter optimization of SVM is a key point in the research of SVM.Fruit fly optimization algorithm(FOA)is an optimization algorithm inspired by the foraging behavior of fruit fly,which seeks the optimal solution of the optimization problem through the cooperation among fruit fly.In the process of searching,FOA has some disadvantages,such as slow convergence and easy to fall into local optimum.Aiming at its shortcomings,this paper proposes the parameter optimization of SVM based on improved FOA.The specific work of this paper is as follows:Aiming at the shortcomings of FOA,a sine-cosine FOA based on chaotic opposition-based learning is proposed.Firstly,in the initial stage of population,the position of fruit fly is evenly distributed in the whole space through chaotic mapping,and elite reverse learning is introduced to improve the quality of initial solution.Secondly,sine and cosine search is introduced in the search process to strengthen global and local search and speed up the transition from global to local.The performance of the proposed algorithm is verified on the test function.Experimental results show that compared with FOA and other similar algorithms,sine-cosine FOA based on chaotic opposition-based learning has better convergence speed and optimization accuracy.Based on the improvement of FOA,the parameter optimization problem of SVM is taken as the problem to be optimized.Through the taste concentration judgment value in FOA,the penalty factor and kernel function parameters of SVM are assigned and optimized,and the optimal parameter combination of SVM is obtained.The performance of the SVM classification model optimized by the improved fruit fly optimization algorithm is verified on UCI data sets.The experimental results show that the proposed classification model has high classification accuracy.Image classification based on support vector machine.Firstly,the image set is preprocessed,then the convolution neural network is used to extract the image features,and then it is sent to the classification model of SVM optimized by improved FOA.The performance of image classification is verified on the image set,and the experimental results show that the proposed method has a good effect in solving the problem of image classification.Figure [27] Table [11] Reference [80]...