Support Vector Machines Optimization Model And Its Application

Support Vector Machine(SVM) is a machine learning method developed on the basis of statistical theory.In particular,its reflected the ideas and methods of structural risk minimization, it can be a good way to solve some practical problems,such as the small samples, nonlinear, high dimension,it also can avoid the local minimum points and the phenomenon of learning, etc.Thus SVM has become a hot research topic in machine learning theory.However, as a new technology,it has yet to be explored and improved in many areas of research.For the problem of small sample characteristics.The predicted results largely determined by the key parameters settings of the algorithm.How to quickly and effectively optimize its argument is a problem of the practical application of SVM.Firstly,this paper were researched and analysis the theory of support vector machine and its related content in the structure optimization design.Secondly,it has described the development status of SVM. studying the fusion technology of improved genetic algorithm and support vector machine. By selecting a Gaussian kernel function and Combining with genetic algorithm to optimize the parameters of support vector machine.Finally, the optimization model was applied to the simulation of grain yield prediction in Hunan Province. With support vector machine has good generalization ability,it can solve the problem which are the complexity of algorithm and input vector dimension closelyrelated issues,thus it can be approximate modeling for any large structure.The research work of this paper includes the following aspects:(1)This paper introduces the research background and significance,also including the organization structure of the support vector machine.(2)It focuses on introduced the theoretical basis of the support vector machine in this paper.Including the statistical learning theory, support vector machine theory, the classical support vector machine and its variant.(3)Discussed and analyzed the problem of choice the kernel function of support vector machine. The kernel function is a key factor When using support vector machine to solve the regression problem,it is a core problem that how to choose an appropriate kernel function for SVM model. Low dimensional space vector sets are tend to be more complex and difficult to be divided,for which you can map them into high-dimensional space which are often easy to divided.However it will increase the computational complexity, it is a good solution that introducted the kernel function. Choosing an appropriate kernel function can greatly reduce the computation in high dimensional space. The most commonly used kernel functions are polynomial kernel function, Gauss radial basis function(RBF) kernel function and sigmoid kernel function.(4)Construct the AGA-SVM model base on adaptive genetic algorithm. In addition to the kernel parameters effect on performance ofsupport vector machines,but also included the penalty coefficient C and other parameters.It Commonly used methods are election of lattice parameter method and experience of the grid point method, But neither of these methods must be verified by a large number of experiments, and the parameters obtained is often not optimal.Followed by the adaptive genetic algorithm to optimized the kernel parameters and other parameters to improve the prediction accuracy.Has low complexity for hardware implementation.(5) Application of AGA-SVM optimization model,By research on the application of support vector machine optimization model, the prediction model was established for Grain yield based on the AGA-SVM.The experiments show that the AGA-SVM model is better than the traditional model,it can accurately reflect the trend of grain yield.