基于改进GA的RBF核函数参数优化模型

RBF kernel function parameter optimization model based on improved GA

为解决高斯径向基(RBF)核函数参数难选择的问题,提出一种基于Fisher判别准则的径向基核函数参数优化模型。为了寻求模型最优解,对传统遗传算法进行了改进,改进的算法保证了适应度大于平均适应度的个体可以遗传到下一代的种群中。通过改进优化求解的效率更高。利用经过改进的遗传算法对所提出的模型进行全局优化求解,最后使用MATLAB 2009a作为工具,并使用UCI数据库的葡萄酒数据作为分析对象进行了实际的算法验证和评价实验。仿真结果表明,提出模型能够有效地找到合适的参数值,提高了SVM的分类准确度。

In order to solve the Gauss radial basis function （RBF） kernel function parameter selection problem, this paper proposes a radial basis kernel function parameter optimization model based on the Fisher criteria. In order to find the optimal solution of the model, the traditional genetic algorithm is improved, and the improved algorithm guarantees the fitness is greater than the average fitness of individuals can genetic to the next generation population. By improving efficiency of the optimization of higher, use an improved genetic algorithm for the proposed model for global optimization solution, finally the use of MATLAB 2009a as a tool, wine data and use UCI database as the object of analysis for algorithm verification and evaluation of practical experiments. The simulation results show that the model can effectively find appropriate parameter values,improve the SVM classification accuracy.