基于概率神经网络的垃圾邮件分类

Classifying and Filtering Spam-mail Based on Probabilistic Neural Network

概率神经网络是由Specht博士在1989年提出的一种径向基神经网络的重要变形。本文提出了把概率神经网络用于垃圾邮件分类,并通过Matlab仿真试验与贝叶斯分类器进行比较,得到了比较理想的结果。

The Probabilistic Neural Network is initially derived from Specht＇ s Modified Radial Basis Neural Network classifier and developed for nonlinear time series analysis . In this paper Probabilistic Neural Networks is applied in spam - mail for classifying and filtering. And the tests result in the simulation experiment proves that the PNN classifier can bring higher precision and efficiency than NaiveBayes Classifier.