面向粒子群优化BP神经网络的粗糙集连续属性离散化算法  

Continuous Attribute Discretization Algorithm of Rough Sets for BP Neural Networks Based on Particle Swarm Optimization

作　　者：毛明扬 徐胜超 MAO Ming-yang;XU Sheng-chao(School of Date Science,Guangzhou Huashang College,Guangzhou 511300,China)

机构地区：[1]广州华商学院数据科学学院,广东广州511300

出　　处：《计算机与现代化》2023年第9期115-119,共5页Computer and Modernization

基　　金：国家自然科学基金面上项目(61772221);广州华商学院校级导师制科研项目(2023HSDS02)。

摘　　要：进行粗糙集连续属性离散化时,获取的断点集并非最优集合,导致离散效果较差。对此,提出面向粒子群优化BP神经网络的粗糙集连续属性离散化算法。对粗糙集连续属性离散化进行分析,采用粒子群算法改进BP神经网络中的权值与阈值,基于优化后的BP深度神经网络对具有连续属性的信息系统进行分类,从而获取多个断点,形成子断点集,进而构建候选断点集。将候选断点集映射成粒子群算法中的粒子,通过改变粒子的速度与位置,找到最佳断点集,完成粗糙集连续属性的离散化。实验结果表明提出的方法可以较好地实现连续属性的离散化,数据一致性最高时趋近于100%,且在不同算法下的分类精度与收敛速度均较高,说明该方法有着较强的应用前景。When discretizing the continuous attributes of rough sets,the obtained breakpoint set is not the optimal set,resulting in poor discretization effect.Therefore,a particle swarm optimization BP neural network oriented discretization algorithm for con⁃tinuous attributes of rough sets is proposed.The discretization of continuous attributes of rough sets is analyzed.Particle swarm optimization algorithm is used to improve the weights and thresholds in BP neural network.Based on the optimized BP deep neu⁃ral network,information systems with continuous attributes are classified,so as to obtain multiple breakpoints,form sub breakpoint sets,and build candidate breakpoint sets.The candidate breakpoint set is mapped to particles in particle swarm opti⁃mization algorithm,and the best breakpoint set is found by changing the speed and position of particles to complete the discretiza⁃tion of continuous attributes of rough set.The experiment results show that the proposed method can better realize the discretiza⁃tion of continuous attributes,and the data consistency is close to 100%when it is the highest,and the classification accuracy and convergence speed are higher under different algorithms,which shows that this method has strong application prospect.

关 键 词：连续属性 粒子群算法 最佳断点集 粗糙集 BP神经网络 离散化 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]