基于全文检索引擎的信息检索技术的应用研究

Research on Full-text Indexing and Searching:an Application Investigation

利用基于JAVA的全文检索引擎Lucene软件包实现了一个电子文档全文检索,并对文本建立索引的4种不同的分词方式进行了讨论及适当地改进。为了提高搜索结果的准确度,在原有文档相关度排序算法的基础上,采用了基于用户行为的二次检索方式。在提高用户体验度方面,采用了Ajax异步通信技术,在用户输入查询条件时,实时地从数据库中提取相关关键词,给与用户提示,供用户选择。

This paper uses a java-based full-text Search Engine software package Lucene to build a document search engine system. The results of 4 different participle methods adopted to build index are compared and an mixed approach is adopted to achieve improvements. It also uses the user＇s behavior to improve the search engine＇s accuracy resorting to an advanced ordering algorithm. Considering of the convenience of the users, the system uses Ajax. When the user input the query, the system will suggest the relevant keywords stored in the database