面向科技文献的多维语义索引构建思路及实现

Construction and Implementation of Multidimensional Semantic Indexing for Scientific Literature

[目的/意义]文章从科技文献检索应用的背景出发,提出了目前检索系统未能满足广大科研工作者的语义检索需求,为了解决这一问题,提出了多维语义索引的新思路。[方法/过程]首先在总结归纳国内外相关研究中主要的语义索引构建方法的基础上得出了要采用的构建方法,然后详细介绍了多维语义索引的设计思路和整体框架,最后以面向物理领域科技文献为例,介绍了其具体实现流程。[结果/结论]从如何针对科技文献内部细粒度知识及其之间的关联关系揭示的角度验证了多维语义索引在语义检索实际应用中的良好效果。

[Purpose/significance]Based on the application of scientific literature retrieval,we found that current scientific literature retrieval system cannot meet the semantic retrieval needs of the majority of scientific researchers.In order to solve this problem,we proposed a new idea for multidimensional semantic indexing of scientific literature.[Method/process]Based on previously prevalent methods of semantic indexing construction,we proposed a new construction method and illustrated the design idea and the overall framework of the multidimensional semantic indexing oriented to the scientific literature.Finally,the specific implementation process was introduced with the physics-oriented scientific literature as an example.[Result/conclusion]The good effect of multidimensional semantic indexing in the practical application of semantic retrieval is verified from the perspective of how to reveal fine-grained knowledge within scientific literature and the association relationships between them.