光盘库搜索引擎的设计与实现-计算机技术专业论文.docxVIP

华中科技大学硕士学位论文 华 中 科 技 大 学 硕 士 学 位 论 文 II II Abstract With the era of big data, enterprise and organizations accumulated lots of important data rapidly. Optical jukebox provides important data technical means for long-term preservation and convenient usage, mainly used in backup and archiving systems. In order to manage and utilize data in a convenient way, these two systems need to be equipped with corresponding retrieval mechanism. Traditional retrieval mechanisms search information according to information added artificially, resulting in low efficiency and low availability. Search engine is often used for the retrieval of information on the Internet, generally using full-text search technology, allowing users to find comprehensive and accurate network resources in a short period of time based on keywords. Introducing search engine into optical jukebox system can provide users with more convenient retrieval services, meanwhile, improve the efficiency of optical jukebox. In order to establish a full-text search system for optical jukebox, the principle of search engine was studied, the architecture of optical disk library system and the property of multi-level storage systems was analyzed, the common retrieval models was estimated. Kinds of text extraction tools were used for extracting content from different types of document; ICTCLAS30 tools was used to complete the function of stemming Chinese word; indexing database was established according to the traditional inverted index algorithm; in order to reduce the frequency of replace discs, a sort method suitable for optical jukebox search system was given; a query retrieval server supporting Boolean query and sort retrieval was accomplished; a client inquiry procedure was implemented, with a user interface designed by VC++6.0 development tool. According to test, the search engine is able to return the correct list of documents which are related to the given keyword, with a higher recall and moderate precision. Moreo