在垃圾邮件分类和朴素贝叶斯算法研究的基础上,提出了基于用户知识的贝叶斯分类算法。通过在分类过程中引入用户知识,克服了电子邮件内容是非结构化、解读依赖于用户的问题。实验证明,面向用户知识的贝叶斯分类算法在商业邮件分类中比普通贝叶斯算法有更好的性能。
An user knowledge based naive bayes classifier was proposed in order to conquer the problem that most of the E- mail is unstructured and need user's decoding. Experiment has proved the method proposed has a better performance then the ordinary naive hayes classifier.