摘要
针对物理世界的信息查找在过去几年间也受到广泛关注,但是迄今还缺乏深入的研究。目前针对Web信息空间的搜索算法不适合普适空间内的信息查询,原因有二:面向物理实体查询的支撑技术,如嵌入式设备和无线通信,与传统Web信息搜索不同;物理实体相关的信息与Web网页不同,表现在元数据、信息动态性等方面。同时,由于用户查询用词与文档关键词不匹配,传统信息检索的结果并不理想,难以满足用户的查询需求。为了解决这个问题,面向物理实体查询这一全新应用,提出了上下文感知矩阵(Context-Awareness Matrix,CAM)的概念,该矩阵表示了本体内上下文之间的关联度,并可根据用户反馈对上下文之间的关联度进行实时训练和动态调整,从而为查询重写和查询扩展提供有效、可靠、稳定的数据支持。实验结果表明该查询扩展算法能获得更合适的查询扩展词。
Information retrieval toward the physical world has been expansively concerned in the past few years ,but still lacks of further research. The query algorithms at present can not be directly used for the information query in ubiquitous space because of the two following differences ：the difference between technologies which are used to support physical entities, and that between the form of information of physical entities and web pages. At the same time, because users＇ query words can＇t always match with the documents,the result of traditional information retrieval doesn＇t seem so satisfying. Therefore, it becomes valuable both in theory and in practice to do more research in query expansion algorithms, so that the problem of ＇No matching words＇ could be minimized. The Context-Awareness Matrix （CAM） is presented to show both the relevance and the similarity between contexts stored in ontology, and the matrix is also able to learn from user＇s feedback and coordinate itself dynamically,which makes it possible to support the system with effective and stable data. The experiment result shows that this query expansion algorithm can make out query expansion words that are more suitable, which could help a lot in obtaining ＇right＇ query results.
出处
《电子设计工程》
2011年第7期14-16,共3页
Electronic Design Engineering
基金
国家高技术研究发展计划(863)重大项目(2009AA011903-1)
关键词
普适计算
查询扩展算法
语义相似度
上下文
本体
ubiquitous computing
query expansion algorithm
semantic similarity
context
ontology
