【摘要】 互联网中数据、信息、知识资源呈现指数级增长,获取这些公开或内部资源的手段分别是传统搜索引擎和站内搜索,这种分离的获取手段造成了信息搜集的不全面,因此对数据融合方法提出了新的挑战。现有的数据融合方法不灵活、集成复杂度高、信息缺失度高。本文提出一种新型的内外数据融合方法,集成自主开发的资源获取组件和成熟的商用服务模块,并通过构建一个应用模型来搭建面向大型机构的学术搜索引擎、形成一个可扩展性强、实时性强、抽取精度高的融合内外部数据的应用平台。该项工作已成功地收集了244个中国科学院所属单位以及相关单位的586,572个网页,34,737个视频,47,390篇论文,并为中国科学院广大师生提供学术资源检索服务功能。更多还原

【Abstract】 The data, information and knowledge resources in internet are growing exponentially. The typical method to obtain those public and internal resources is reached by querying internet search engine and custom search engine respectively. It is not only difficult but also incomplete in the resulted information collection. Therefore, it raised a new challenge for a good performance data fusion method. The existing methods are not flexible, and may cause high integration complexity and high information scarcity. This paper proposes a new kind of internal and external public data fusion method. We developed a new resources access component added to a mature commercial service module and applied to constructing an academic search engine for large institutions, and implemented a search service platform which can get both external public content and internal information. It shows good extensibility, real-time performance and high accuracy. The platform has successfully collected 586,572 web pages, 34,737 videos and 47,390 papers from 244 subordinate departments in CAS and related units, and providing academic retrieval service to the faculty and students at Chinese academy of sciences.更多还原