摘要
基于贝叶斯模型,提出了一种更为灵活的、有效的、基于全局的即时分类模型的垃圾邮件分类模型,pValue优先邮件过滤模型。该分类模型着眼于全局的实例,在所有邮件中选取最需要得到评估的实例进行计算,反复进行该过程并在需要的时候中断进程,最终得到实例的全局最优化。我们使用UCI提供的邮件样本进行验证。实验结果验证了该分类模型比FCFS即时分类模型的性能更优。在即时分类的早期较好的改善分类效果降低分类的0-1损失错误率和RMSE错误率,并伴随着计算资源的增加得到更好的分类准确率。
This paper proposes a new anti-spare model： p-Value first anti-spam model. This model addresses on the whole mail, and lets email which needs get evaluation first get new calculate resource. It can get the best result on whole entry. The UCI＇s email sample has been used to examine this model. The result of experiment shows that this model can get better performance.
出处
《电子测量与仪器学报》
CSCD
2009年第5期46-51,共6页
Journal of Electronic Measurement and Instrumentation
关键词
贝叶斯分类
即时分类
垃圾邮件过滤
Bayes classification
anytime classification
anti-spam
