基于果蝇优化的支持向量机回归模型

摘要

给出一种基于果蝇优化的支持向量机回归模型。将支持向量机惩罚因子和核函数参数初始化为果蝇群体，根据果蝇优化算法原理，依据适应度最优原则进行迭代觅食，搜索最优参数，建立模型。将该模型用于分析有机化合物熔点预测问题，结果显示，该模型预测均方误差为3．02％，相关系数达到89．39％。%A support vector machine (SVM ) regression model based on fruit fly optimization algorithm is proposed .Initialize the SVM penalty factor and kernel function parameter as a fruit fly group .By the rule of fruit fly optimization ,execute iterative optimal foraging according to the fitness principle until the optimal parameters are sought out ,thus ,the model can be set up .Use this model to analyze the melting point prediction of organic compounds ,it turns out that ,the prediction error is 3 .02% , and the correlation coefficient is 89 .39% .