测试样本空间变化对贝页斯常规及补集规则权重评估影响的分析
Impact of the Change of Training Sample Space on Bayesian Regular and Complement Class Rules Weight Estimate
-
摘要: 为降低防火墙文本分类计算的误码率,研究基于贝页斯模型的防火墙测试系统的运行效率,提出将贝页斯模型视为线性模型的观点,分析测试样本空间变化对模型不同集合规则权重的影响,建立了有误差补偿功能的MNB分类器数学模型,实验仿真验证了贝页斯多项式数学模型的可行性,确定了MNB分类器内贝页斯多项式数学模型特征变量与目标文本内部个性词汇的对应关系.Abstract: In order to reduce the error rate in classification calculation of the firewall text, the operation efficiency of the Bayesian firewall test system was studied, and a new viewpoint was proposed to set Bayesian model as a linear model. To analyze the influence of test sample space changes on the different set weights rules,a MNB classifier mathematical model was set up with error compensation function. The feasibility of Bayesian s polynomial mathematical model was verified by experimental simulation,and the corresponding relationship was determined between characteristic variables of Bayesian s polynomial mathematical model in MNB classifier and character vocabulary in target text.
下载: