摘要
研究了一种基于动态神经网络支持向量机(SVM)的FPGA硬件实现方法.提出了基于动态神经网络的最小二乘支持向量机(LS-SVM)神经网络结构,完成了VHDL语言描述的基于动态神经网络的LS-SVM结构设计,并在XILINX SPANT3E系列FPGA中完成了LS-SVM的分类与回归实验.结果表明,该硬件实现方法很好地完成了SVM的分类与回归功能,与现有的软件仿真和模拟器件实现相比,该方法具有更快的收敛速度和更高的灵活性.
A new FPGA hardware implementation approach of dynamic neural network for support vector machines was provided and researched.The structure of dynamic neural network for least square support vector machines(LS-SVM) was proposed.The architecture design of dynamic neural network for LS-SVM based on VHDL language was also performed.The experiments of classification and regression for LS-SVM were achieved on XILINX SPANT3E series FPGA.The experimental results show that it is effective to complete the LS-SVM classification and regression based on presented method.Compared with the(existing) methods based on software implementation or analog device implementation,this approach has(better) convergence rate and better flexibility.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2010年第7期962-967,共6页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金(60675048)
陕西省自然科学基金(2007F30)资助项目
关键词
支持向量机
最小二乘支持向量机
动态神经网络
稳定性
support vector machines(SVM)
least square support vector machines(LS-SVM)
dynamic neural network
stability