Lucene的最小风险概率加权朴素贝叶斯算法

摘要

In order to improve the accuracy of spam filtering and analyze the spam data, on the basis of the comparison between information retrieval and information filtering, the relationship between the information retrieval Lucene application framework to spam filter system is put forward, and minimum risk probability weighting simple Bayesian algorithm is combined with minimal risk, effectively reducing the Bayesian method of independence constraint. Experimental results verify that the weighted naive Bayes algorithm is effective.%为了提高垃圾邮件过滤的准确性,在分析垃圾邮件数据的基础上,对比信息检索与信息过滤之间的关系,将信息检索框架Lucene应用到垃圾邮件过滤系统中,提出最小风险概率加权的朴素贝叶斯算法,与最小风险法结合,有效地减少贝叶斯方法的独立性约束.实验验证了加权朴素贝叶斯算法的有效性.