摘要
在检验知识较少的情况下,为有效提取网络数据特征,实现快速入侵检测,运用边界检测算法建立有限样本下自学习的理论框架和通用方法,通过特征提取来剔除对其后的入侵检测不具有鉴别信息的部分。实验数据分析表明,有效聚类的数量随着原始样本量的增加而缓慢增加,并趋于稳定。理论分析和仿真实验证明了该算法的有效性。
Adopting the boundary detection method can help the users achieve an effective extrac- tion of the network data and a quick intruding detection in the lack of detective knowledge. It is help- ful to establish a theoretical framework and develop a universal method of self-learning as to the limit- ed samples with the boundary detection method, through feature extraction can get rid of the data without the distinguished features in the following intruding detection. Experimental data analysis shows that the number of effective clustering increases along with the increasing original samples and then levels off. Its validity has been demonstrated in the theoretical analysis and simulation experi- ments in this article.
出处
《黑龙江科技学院学报》
CAS
2011年第2期142-145,共4页
Journal of Heilongjiang Institute of Science and Technology
基金
黑龙江省教育厅科学技术研究项目(11551439)
关键词
边界检测算法
入侵模式
特征提取
boundary detection algorithm
invasion pattern
feature extraction