Both-branch fuzzy decision and decision encryption-authentication

This paper is on fuzzy decision theory and information security theory which are mutually independent and engrafted. This paper presents both-branch fuzzy decision and problems on decision encryption-authentication, puts forward two kinds of both-branch fuzzy decision on X: both-branch fuzzy decision on X having bounded domains X = X ∩X- ={x0}, both-branch fuzzy decision on X havingoverlapping domains X* = X ∩X- = {x_1~*,x_2~*,…,x_1~*}. It presents single branch separate-separate out theoremof both-branch fuzzy decision on X, single branch superposition theorem of both-branch fuzzy decision, exclusive-decompose principle of both-branch fuzzy decision. This paper presents encryption-authentication theorem of decision by engrafting information security theory and technique with both-branch fuzzy decision, and provides encryption-decryption and signature-authentication of both-branch fuzzy decision.     

一种求解优化问题的新型混合遗传算法

论文在标准遗传算法中引入新的交叉运算和变异运算,针对优化问题提出了一类新型混合遗传算法。具体算例验证了算法的有效性和相对于标准遗传算法及某些混合遗传算法的优越性。     

基于投影寻踪和带惩罚函数的SVM的入侵检测分析

对于网络入侵检测问题,本文考虑先将原数据进行投影寻踪降维,而后对降维的数据利用带罚函数的支持向量机方法进行分类识别。利用模型通过对KDD99数据集的计算及检验,效果较为理想。     

基于相似属性PCA和SVM的网络入侵检测

针对现有网络入侵检测算法泛化能力差与处理大样本数据耗时长的问题,本文提出了基于相似属性主成分分析（PCA）与支持向量机（SVM）的网络入侵检测的方法。采用KDD 1999数据集仿真,原始数据集根据属性间的相似程度分为四类属性集,对四类属性集分别采用PCA进行特征抽取,最后用SVM检验分类的正确率。实验结果表明：与直接采用PCA对全部属性一起降维相比,相似属性PCA的分类降维方法有较短的处理时间,并且有更强的泛化能力,即对未知攻击类型的检测性能。