文献类型：期刊文献

中文题名：基于改进粒子群算法的孪生支持向量机

英文题名：Twin support vector machine based on improved particle swarm optimization algorithm

作者：顾吉峰[1];王蓓[1]

机构：[1]华东理工大学化工过程先进控制和优化技术教育部重点实验室,上海200237

年份：2020

卷号：41

期号：11

起止页码：3078

中文期刊名：计算机工程与设计

外文期刊名：Computer Engineering and Design

收录：CSTPCD;;北大核心:【北大核心2017】；

基金：国家自然科学基金项目(61773164);上海市自然科学基金项目(16ZR1407500)。

语种：中文

中文关键词：粒子群搜索算法;适应值增益;渐变扰动;孪生支持向量机;参数寻优

外文关键词：PSO;adaptive value gain;gradual disturbance;twin support vector machine;parameter optimization

摘要：为解决粒子群搜索算法局部最优解和收敛效率低的问题,提出一种改进型的粒子群搜索算法(IPSO)。为速度惯性权重引入自适应增益反馈率,提高收敛速度;引入渐变随机扰动,利用局部不确定性,跳出局部最小;利用IPSO对TWSVM的参数实现寻优。4种基准函数对IPSO的搜索性能的分析结果表明,IPSO有着更好的搜索能力和收敛速度,IPSO-TWSVM在不同数据集分类中的收敛速度和分类准确率上均优于其它算法。
To solve the local optimal solution and low convergence efficiency of the particle swarm search algorithm,an improved particle swarm search algorithm was proposed.The adaptive gain feedback rate was introduced for the speed inertia weight and the convergence speed was improved.The gradient random disturbance was considered and the local uncertainty was utilized to jump out the local minimum.The developed IPSO algorithm was adopted to optimize the parameters of the twin support vector machine.The comparison results of the four benchmark functions show that the proposed IPSO has better search ability and convergence speed.The classification performance of IPSO-TWSVM is better than that of the other algorithms in convergence speed and classification accuracy.