Abstract: A NLG（Natural Language Generation） based XML keyword search technology is presented in this paper，to resolve the problem of losing semantics for XML keyword search based on LCA（Lower Common Ancestor）.The document planner of NLG is applied to XML document and consequently message collection is deduced.With the user filtering the message collection，the semantic search is achieved and the efficiency is improved greatly.

摘要： 为了解决基于LCA（Lower Common Ancestor）的XML关键字查询丢失语义的问题，提出了一种基于“自然语言生成技术（Natural Language Generation，NLG）”的XML关键字查询技术，将NLG的内容规划应用到XML文档，产生针对用户查询的消息语句集，通过对消息语句集的筛选既可以实现基于语义的XML关键字查询，又可以极大地提高查询效率。