【摘要】 为了提高多载体图书信息检索能力,提出多载体图书信息邻近似目标精细化检索系统设计方案。采用分位数回归模型构建多载体图书信息筛选模型;采用关联信息熵特征提取方法进行多载体图书信息的近邻特征挖掘,提取多载体图书信息的关联语义特征量;利用语义本体特征重构方法进行多载体图书信息的检索和筛选识别,分析多载体图书信息的输出高维特征信息;采用邻近相关性检索方法实现多载体图书信息的语义信息检索和自适应查询,实现多载体图书信息的信息分类和特征匹配。在算法设计基础上,采用嵌入式的Linux技术进行多载体图书信息邻近似目标精细化检索系统的开发设计。测试结果表明,采用该方法进行多载体图书信息邻近似目标精细化检索的召回性较好,检索效率较高。更多还原

【Abstract】 A design scheme of the proximity-like target refined retrieval system for multi-carrier book information is proposed to improve the multi-carrier book information retrieval ability. The quartile regression model is used to construct the multicarrier book information screening model. The relevance information entropy feature extraction method is used to mine the nearest neighbor features of multi-carrier book information,so as to extract the associated semantic feature quantity of multi-carrier book information. The semantic ontology feature reconstruction method is used to retrieve,screen,and recognize multi-carrier book information,so as to analyze the output high-dimensional feature information of multi-carrier book information. The adjacent relevance retrieval method is used to realize semantic information retrieval,adaptive query,classification and feature matching of multi-carrier book information. On the basis of the algorithm design,the embedded Linux technology is used to develop and design the proximity-like target refined retrieval system for multi-carrier book information. The test results show that the method has a good recall performance and high retrieval efficiency for proximity-like target refined retrieval of multi-carrier book information.更多还原