基于果蝇优化的SVM回归模型及其应用研究  

SVM Regression Model Based on Fruit Fly Optimization and Its Application

作　　者：吴悦[1] 

机构地区：[1]淮阴工学院现代教育技术中心,江苏淮安223003

出　　处：《淮阴工学院学报》2015年第1期18-20,40,共4页Journal of Huaiyin Institute of Technology

摘　　要：为了有效解决支持向量机模型在参数选择上的盲目性,提高该模型的学习性能和泛化能力,提出一种基于果蝇优化的SVM方法。该方法首先运用果蝇优化算法选择全局最优的SVM惩罚因子和核函数参数,建立SVM分类模型,进而将该模型应用于对有机化合物的熔点预测问题中。实验结果表明,基于果蝇优化的SVM模型效率高,实际应用效果好。In order to effecti rely solve the problem of blindness of support vector machine （ SVM ） model on the parameter selection, improve the learning performance and generalization ability of the model, a SVM based on fruit fly optimization method was put forward. Drosophila optimization was first used as an optimization algorithm to select the global optimal SVM penalty factor and the kernel function parameter so as to establish the SVM classification model, which was then used to simulate the actual problem. The model was applied to predict melting point of organic compcunds. The experimental results showed that the SVM based on fruit fly optimization model had high efficiency and good effect in practical application.

关 键 词：支持向量机 果蝇 有机化合物 熔点 预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]