A Novel Semantic Association Mining Method

Abstract

The invention discloses a novel semantic association digging method comprising the following steps of: analyzing and clustering input semantic data to form a semantic data cluster, and cleaning data to convert a resource description frame diagram into a type object diagram; blocking or combining the type object diagram to obtain a plurality of blocks; predicting a potential link manner and a semantic association order of magnitude in each block, feeding the predicting result back to a second-division unit, and further dividing bigger or complex blocks; digging a local link manner and semantic associations of the blocks, summarizing and counting the semantic associations, and outputting the result to users. According to the above manner, the novel semantic association digging method provided by the invention has the characteristics of high efficiency, accurate digging result and the like; complex associations among a plurality of objects can be described, and the link manner is used for judging the typicality of semantic associations; and a diagram digging technology is used for digging so that the digging feasibility and efficiency of semantic associations on large-scale semantic data are improved.

Description

technical field [0001] The invention relates to the field of information retrieval, in particular to a novel semantic association mining method. Background technique [0002] With the vigorous development of the Semantic Web in the past ten years, online semantic data has become more and more abundant, and huge semantic data sets constitute a complex data network. Semantic search in the Semantic Web mainly focuses on semantic objects and semantic associations between objects. The goal of semantic association retrieval is to help users find and understand direct or indirect connections between objects hidden in massive semantic data. [0003] In the field of Semantic Web research, semantic association is usually defined as the direct or indirect relationship between objects in the resource description framework. The modeling of semantic association usually follows the directed path in graph theory. For given two objects, the process of semantic association discovery is to qu...

AI Technical Summary

Problems solved by technology

The traditional semantic association model based on semantic path still has limitations: (1) The traditional semantic path can only describe the semantic association between two objects, and each semantic association is independent of each other, cannot be unified, and cannot describe the semantic association between multiple objects. Complex associations, but in fact a large number of objects exist in real semantic data, which should be unified into a whole semantic association; (2) The semantic path model does not consider the typicality of semantic associations, that is, two objects with semantic associations Whether the semantic path between them also appears in other semantic associations. In many cases, the trivial and unimportant semantic path only indicates the connectivity of two objects in the resource description framework diagram, and does not indicate the relationship between two objects. There is a meaningful semantic relationship between

The PartMiner algorithm is the most popular graph partitioning algorithm for graph mining. However, the algorithm has the possibility of mining wrong results in theory, and it needs to further check the correctness of the mining results after global mining. Therefore, there is no perfect partitioning algorithm so far. The block method can quickly and accurately divide and merge large-scale semantic data

Images