基于LDA和本体的科学前沿识别与分析方法研究

Research on the Method of Detecting and Analyzing Scientific Fronts Based on LDA and Ontology

[目的/意义]鉴于及时准确把握科学前沿的重要意义,文章针对目前科学前沿识别方法存在的问题,提出基于LDA和本体的科学前沿识别及分析方法。[方法/过程]通过LDA模型抽取研究主题,采用主题强度和主题新颖度两个指标来识别科学前沿主题,并基于领域本体进行概念映射来挖掘科学前沿主题的语义类型,从而实现科学前沿的语义分析。[结果/结论]基于LDA模型抽取的10个研究主题中,确定了4个科学前沿。该方法在科学前沿主题的表示方面,主题识别的方法和结果的语义分析方面都有一定的改进。

[Purpose/significance] Given the necessity of timely and accurate identification of scientific fronts,this paper proposes a method of scientific fronts detection and semantic analysis based on LDA model and domain ontology. [Method/process]LDA model is applied to extract research topics form documents,and topic intensity and topic novelty are used as indicators to recognize scientific front topics. To realize the semantic analysis of research fronts,domain ontology is used to map the semantic types of topics. [Result/conclusion] Four research fronts are obtained form the 10 extracted research topics based on LDA. This method can effectively improve the performance of the topic representation,recognition and semantic analysis.