【摘要】 分析了一种基于直线几何分割的朴素贝叶斯邮件过滤模型LGDNBF,用更为精确的代价因子描述了分类器误判的代价。定义了高风险决策区域,对高风险决策区域中的邮件引入SVM方法进行二次分类,提出了基于精确代价因子的两层邮件过滤模型。在中文邮件语料集上的实验结果证明了这一两层过滤模型的分类效果较之朴素贝叶斯邮件过滤模型有明显的改进。更多还原

【Abstract】 A new Naive Bayes Filtering model based on Line Geometry Division(LGDNBF)is analyzed.Cost of false positive classification is described by smarter cost factor.With definition of the high risk area,the paper introduces SVM to classify second time,and puts forward a bi-layer email filtering model based on smart cost factor.The test on Chinese email corpus demonstrates that the new model has better performance than Naive Bayes Filtering model.更多还原