摘要
近年来,随着计算机和数据库技术的快速发展,大规模数据集迅速增长,利用特征选择技术来筛选信息量大的特征已经变得非常重要。本文提出了一种具有混合策略的樽海鞘群特征选择算法(salp swarm feature selection algorithm with hybrid strategy,HS-SSA)。首先,本文生成一张基于互信息的排序表,并由排序表提出了新的初始化策略。其次,提出一个新颖的并且有条件调用的动态搜索算法。最后在位置更新上结合瞬态搜索算法(transient search algorithm,TSO),改进勘探和开发步骤的效率,增加解空间的灵活性和多样性,从而使算法能够快速定位到全局最优位置。为了验证算法的性能,实验选取14个UCI的数据集,并且与樽海鞘群算法(SSA)以及近几年樽海鞘群的改进算法等多种优化算法进行比较,结果表明HS-SSA在特征选择上具有更强的竞争力。
In recent years,with the rapid development of computer and database technologies,the number of large-scale datasets has rapidly increased.Thus,the use of feature selection technology is important to screen features with massive amounts of information.In this study,a salp swarm feature selection algorithm with a hybrid strategy(HS-SSA)is proposed.Initially,a sorted table based on mutual information is generated,and a new initialization strategy is proposed on the basis of this sorted table.Furthermore,a novel dynamic search algorithm with conditional call is proposed.With respect to location updates,the efficiency of exploration and development steps is improved,and the flexibility and diversity of the solution space are increased by combining the transient search algorithm(TSO).Consequently,the algorithm can rapidly locate the global optimal location.To verify algorithm performance,14 UCI datasets were selected for the test.In addition,the proposed algorithm was compared with the salp swarm algorithm(SSA),the improved SSA,and many other improved algorithms in recent years.The results show that HS-SSA is more competitive in feature selection.
作者
余紫康
董红斌
YU Zikang;DONG Hongbin(School of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China)
出处
《智能系统学报》
CSCD
北大核心
2024年第3期757-765,共9页
CAAI Transactions on Intelligent Systems
基金
黑龙江自然科学基金项目(LH2020F023).
关键词
特征选择
樽海鞘群算法
瞬态搜索算法
启发式算法
互信息
动态搜索算法
秩和检验
K近邻
feature selection
salp swarm algorithm
transient search algorithm
heuristic algorithm
mutual information
dynamic search algorithm
rank sum test
K-nearest neighbor