Abstract: Traditional topic-specific search is difficult to meet the needs of a wide range of vertical intelligent search. Topic-specific search and semantic search technologies are applied to the data collection and intelligent inquiry process respectively. Specifically, this paper uses the domain data collection robot based on hierarchical classification model to gather the domain information. Besides, based on Chinese Encyclopedia Resource, a domain-ontology is built automatically, and is applied into semantic inference and query expansion of search engine, so a domain-oriented intelligent search engine is proposed in order to achieve professional and intelligent search. The results show that, the hierarchical classification for domain has a more high precision and recall?rate, and this system has many advantages compared?with?other search?engines, such as highly specialized domain, easy to?transplant, more?intelligent?and so on.

Key words: topic-specific search, domain data collection, semantic search, domain ontology, hierarchical classification

摘要： 传统的主题搜索技术难以适应大范围垂直领域的智能搜索需求，通过将主题搜索与语义搜索相关技术分别应用到搜索引擎的数据采集与智能查询过程中，利用基于层次分类模型的领域数据采集机器人，完成对领域信息的精准采集，基于中文百科资源自动构建领域本体，将大规模领域本体库用于搜索引擎的语义扩展推理中，实现了一个面向领域的智能搜索引擎。实验结果表明，基于层次结构的领域分类具有较高的分类准确率和召回率，与其他搜索引擎相比较，该系统具有领域专业性强、领域易于移植、检索更加智能等特点。

关键词: 主题搜索, 领域数据采集, 语义搜索, 领域本体, 层次分类