基于综合语义相似度的化工信息检索方法

打开文本图片集

中图分类号：TQ011;G252 文献标志码：A 文章编号：1001-5922（2025）07-0121-04

Abstract：To improve theaccuracyof chemical informationretrieval，acomprehensivesemantic similarityretrieval method was proposed.A comprehensive semantic similarity calculation model was constructed by combining string semantic similarity，node semantic distance similarity，and neighboring nodesimilarity.To determine the weightsof diffrent semantic similarities，a generalized regression neural network（GRNN）optimized bythebat algorithm was used to train and alocate weight values.The comprehensive semantic similarity formula was used to retrieve semanticsinthe fieldof chemical engineering.Theresults indicatethatthesemanticretrieval resultsofchemical information obtained bythe proposed method are close to the expert rating results，with an average Pearson corelation coefficient value ofO.94.Compared to the comparison method，the calculation results of this method are more similar to the expert rating results.

Keywords：comprehensive semanticsimilarity;chemical information;semanticretrieval;bat algorithm；GRNN network

信息检索是利用标准化知识表达对历史数据进行复用，为安全分析提供有效信息的重要途径。（剩余5660字）