基于概念格的情景范围实例检索算法

Concept-lattice-based Context Scope Instance Retrieval Algorithm

历史版本的情景可以转化为知识实例,用于后续查询和推理。知识查询需要涉及到实例检索过程。实例检索问题可以通过ABox满足性测试(ASAT计算)完成,ASAT计算可通过Tableau算法实现。此算法比较耗时。对此,已经存在一系列技术用于优化实例检索的时间性能。该文进一步关注具有情景范围的情景知识实例检索。为了提高这一类实例检索的时间性能,提出使用概念格索引情景知识实例。由于每次查询范围不可能完全等价于格内涵所代表的区间,提出了基于概念格的实例检索算法实现特定范围的查询。通过实验和原始查询比较,使用概念格索引和此算法,时间性能得到了提高。

History context can be converted to knowledge instances for context query and reasoning. Knowledge query is referred to instance retrieval process. Instance retrieval can be completed by ABox satisfy test （ASAT computation）, ASAT computation can be implemented by Tableau algorithm. This algorithm is time-consumed. Thus, there have been a series of optimized instance retrieval techniques to improve the time performance. This paper further focuses on the context knowledge instance retrieval within context scopes. To improve the time performance of this sort of instance retrievals, this paper proposes using lattice indexing context know- ledge instance. For each query scope cannot be equal to the intervals expressed by lattice intents exactly, this paper proposes a con- cept-lattice-based instance retrieval algorithm to implement the queries with the given scopes. Compared with the normal query way, the time performance is improved by using this algorithm with concept lattice index.