摘要
在反垃圾邮件系统中,基于内容的过滤技术得到了广泛应用,将贝叶斯统计方法应用于邮件文本过滤取得了较好的性能。本文利用贝叶斯过滤算法构建了反垃圾邮件系统模型,应用系统性能评价指标对贝叶斯过滤算法的多变量贝努利事件模型和基于词频的多项式事件模型进行验证及评价,实验结果表明,若保证有充足的训练样本条件下,并根据Spam的变异调整训练样本集,贝叶斯过滤算法会获得较高的Spam过滤性能,反垃圾邮件效果显著。
In the anti -spare systems, content- based filtering technology has been widely used and. Bayesian statistical method is used in the message text filter to obtain a better performance. In this paper, Bayesian filtering algorithm is applied to construct the anti - spam system model. The application performance evaluation of Bayesian filtering algorithm multivariate Bernoulli event model and e- vent model based on word frequency polynomial verification and evaluation of experimental results indicated that if suffleient training sample conditions, the training sample set and in accordance with the variation adjustment of spam, Bayesian filtering algorithm will be high spare filtering performance and anti - spare results are obvious.
出处
《内蒙古农业大学学报（自然科学版）》
CAS
北大核心
2013年第3期163-167,共5页
Journal of Inner Mongolia Agricultural University(Natural Science Edition)
基金
内蒙古工业大学基金项目(X200801)
