摘要
针对网络入侵检测所处理数据存在特征维数高、检测效率低、准确率不高的问题,提出一种改进麻雀搜索算法的特征选择方法,旨在减少特征冗余的同时提高分类准确率。利用改进Circle映射初始化种群;结合秃鹰搜索算法中的螺旋搜索方式更新发现者位置;采用单纯形法和小孔成像法优化适应度较差和最优麻雀的位置,提升算法的寻优能力。将该算法与其它算法在6个经典基准函数上进行对比测试,其在收敛速度、精度等方面均有提升。使用数据集CIC-IDS2017进行特征选择实验,平均保留了7.6个特征,准确率达到了99.5%,结果表明,该算法可以在保证准确率的同时有效降低特征维度。
Aiming at the problems of high feature dimension,low detection efficiency and low accuracy of the data processed by network intrusion detection,a feature selection method based on improved sparrow search algorithm was proposed to reduce feature redundancy and improve classification accuracy.The improved Circle map was used to initialize the population.The location of the discoverer was updated by combining the spiral search method in the vulture search algorithm.The simplex method and small hole imaging method were used to optimize the location of the poor fitness and the optimal sparrow,to improve the optimization ability of the algorithm.The algorithm was compared with other algorithms on six classical benchmark functions,and its convergence speed and accuracy were improved.The feature selection experiment was carried out using the dataset CIC-IDS2017,which retained an average of 7.6 features with an accuracy rate of 99.5%.The results show that the algorithm can effectively reduce feature dimensions while ensuring the accuracy.
作者
刘涛
蒙学强
LIU Tao;MENG Xue-qiang(College of Communication and Information Technology,Xi’an University of Science and Technology,Xi’an 710600,China;Xi’an Key Laboratory of Heterogeneous Network Convergence Communication,Xi’an University of Science and Technology,Xi’an 710600,China)
出处
《计算机工程与设计》
北大核心
2024年第4期989-996,共8页
Computer Engineering and Design
基金
国家自然科学基金项目(61674121)。
