以关联开放数据为代表的关联数据将越来越多的出现在互联网中，因此面向关联数据进行信息检索的研究具有现实意义。本项目将解决面向关联数据进行信息检索的关键技术问题。针对自然语言查询接口问题，分析各种查询类型，识别查询意图，结构化表示问题语义，然后将结构化问题语义与知识库中的资源形成映射，进而完成自然语言-SPARQL语句之间的转换；针对联合查询的效率问题，基于分布式结构化本体数据源，研究数据源选择方法、联合查询优化和执行策略，进而构建联合查询优化模型；针对查询结果的本体数据源排序问题，以多种本体排序算法的结果作为特征，以这些特征为数据源使用各种排序学习算法进行排序学习，使用机器学习的方法优化各种排序算法和排序学习算法的权重，进而构建本体排序通用模型。本项目将对基于关联数据的精准检索和语义网技术的发展具有重要意义。

Linked data will be more and more represented by the Internet as Linked Open Data, thus information retrieval from Linked Data has practical significance. This research will solve the key technology problems of information retrieval based on Linked Data. Aiming at the natural language query interface, many natural language query types are analyzed, and their query intents are identified, and their semantic structures are expressed, then the mappings between the query semantic structures and the resources of knowledge base are formed, lastly the translation of natural language query to SPARQL is completed. Aiming at the efficiency of joint query, data source selection method and joint query optimization and execution strategies are researched based on distributed ontology data, and then the joint query optimization model is constructed and completed. Aiming at the ontology data source ranking of query results, the results of various ontology ranking algorithms are as features, then ‘learning to rank’ is implemented by various ranking learning algorithms based on these features, and the machine learning approaches are used to optimize the weights of these features and ranking learning algorithms, lastly a general model of ontology ranking is constructed. On the basis of the above research, an information retrieval prototype system based on Linked Data is constructed. The research will be of great significance on the accurate retrieval based on Linked Data and the development of semantic web technology .