Improved Support Vector Machine Based On Sparrow Search Algorithm And Its Applicatio

The support vector machine algorithm is based on structural risk minimization and statistical learning theory,which is widely used in pattern recognition and has good generalization ability,among which how to choose the appropriate kernel function and kernel parameters is an important part of the algorithm.The sparrow search algorithm simulates the sparrow’s strategy of avoiding predators in the process of searching for food,with better search structure,iteration and global exploration ability,and the search results have more stability.However,the sparrow search algorithm also has some shortcomings,such as the search area tends to be the origin,and the iterative formula depends on the value of the objective function.Therefore,this paper improves the sparrow search algorithm and combines the improved sparrow search algorithm with multi-kernel support vector machine to achieve better classification performance.The research of this paper includes the following two aspects:1.To address the problems of the sparrow search algorithm,such as the lack of population diversity and the tendency to fall into local optimum in the late iteration,we improve the algorithm by incorporating Bernoulli mapping,opposition-based learning,firefly algorithm and Gaussian variation,and conduct simulation experiments using test functions.The experimental results show that the improved sparrow search algorithm has better performance compared with other optimization algorithms.2.In order to improve the classification performance of support vector machines,a multi-kernel support vector machine is constructed using linear and Gaussian kernels as the basis kernel functions.The improved sparrow search algorithm is used to search for multi-kernel support vector machine parameters,and training and prediction are conducted on transformer fault instances and air quality level classification.The experimental results show that compared with other diagnostic models,the multi kernel support vector machine method optimized based on the improved sparrow search algorithm has higher classification accuracy.