中文摘要：

为实现对网络敏感信息的检测和过滤,提出一种基于确定有穷自动机的改进算法ST-DFA（swift tree DFA）。对传统的DFA过滤算法进行改进,不再依赖敏感信息语料库,只须建立一次敏感信息决策树,即可实现对网络信息的多次过滤;当敏感词语料库发生更新时,可以实现对敏感词的决策树的实时更新。实验结果表明,ST-DFA算法有较高的工作效率,适合当下对互联网敏感信息的检测与过滤。

英文摘要：

To realize the detection and filtration of sensitive network information,based on the improved algorithm of finite automaton,ST-DFA（swift tree DFA）algorithm was proposed.Traditional DFA filtering algorithm was improved.ST-DFA method no longer depended on sensitive information corpus,only a decision tree needed to be built,and multiple sensitive network information filtering was realized.When the words corpus was updated,real-time updating of decision tree of sensitive words was realized.Experimental results show that ST-DFA algorithm has better working efficiency,and it is suitable for the sensitive network information detection and filtration.