摘要
该文研究了当训练点的输出为三角模糊数时,支持向量回归机的构建问题。首先将模糊回归问题转化为模糊分类问题,并将求模糊最优分类超平面问题转化为求解带有模糊决策的机会约束规划问题。利用基于模糊模拟的遗传算法求解带有模糊决策的机会约束规划,得到模糊最优分类超平面(模糊方程),解模糊方程得到模糊回归函数。在此基础上,得出模糊线性支持向量回归机(算法)。从而较好地解决了支持向量机中含有模糊信息的模糊回归问题。最后,给出显示模糊线性支持向量回归机特点的模糊支持向量集的定义。
This paper focuses on the construction of Support Vector Regression when the outputs of training points are triangle fuzzy numbers.First this paper transforms fuzzy regression problem to fuzzy classification problem,then transforms it to chance constrained programming with fuzzy decision,and solves this programming using fuzzy simulation based evolutionary algorithm to get the optimal classification hyper-plane(fuzzy equation),at last solves the fuzzy equation to derive the fuzzy regression function.Based on this,the paper proposes the Fuzzy Linear Support Vector Regression (FLSVR).This can deal with regression problem with fuzzy information well.At the end,the definition of fuzzy support vector set is given,which can declare the characters of FLSVR.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第36期54-57,共4页
Computer Engineering and Applications
基金
国家自然科学基金资助项目(编号:10371131)
关键词
模糊规划
模糊回归
模糊模拟
模糊数
fuzzy programming,fuzzy regression,fuzzy simulation,fuzzy numbers
