An optimized spam filtering method based on AODE

作者：Chen, Bin; Dong, Shoubin^*

来源：Journal of Information and Computational Science, 2009, 6(2): 749-756.

• Averaged one-dependence estimators
• Data set
• Improvement technique
• Machine learning methods
• Naive bayes
• Spam filtering
• Spam filtering technique
• Training optimizations

With the development of spam filtering techniques, many machine-learning methods are introduced to solve spam problems. Among all these methods, Naive Bayes is widely used in Spam detecting due to its simplicity and efficiency. However Naive Bayes is based on the assumption of attribute independence and this assumption is not fit for the reality and may constrain the accuracy. This paper firstly introduced the state-of-arts of the improvement techniques to weaken this assumption, and then proposed a method based on ADOE (Averaged One-Dependence Estimators). This method has been implemented on our developed filter by making some training optimization. The experimental results on standard data set show that this improvement can enhance the performance of the Naive Bayes filter.

• 出版日期2009-4
• 单位华南理工大学