关键词: 论文创新性, 研究主题, 情报学, 隐含狄利克雷分布（LDA）, 支持向量机（SVM）

Abstract: [Purpose/Significance] Theme innovation is one of the most essential features of the innovation of scientific papers.This paper aims to analyze the innovativeness of information science papers based on the perspective of theme evolution, so as to provide a new insight for dynamic evaluation. [Design/Methodology] Papers selected from 11 CSSCI journals in the field of information science of the past 20 years are as samples, a method combined the LDA（Latent Dirichlet Allocation）topic model with SVM（Support Vector Machine）classification algorithm is utilized to identify potential topics in abstracts and judge the innovation of papers. Finally, statistical methods are used to verify the accuracy of the evaluation results. [Findings/Conclusion] The evaluation method of academic papers applied in this study can effectively identify papers with innovative value in different periods in the field of information science, and can provide references for selection of research topic,evaluation of paper themes innovation, and review of journal papers. [Originality/Value] This study expands the application field of innovative evaluation methods integrating LDA and SVM, and enriches the innovative evaluation system of scientific research papers based on content. 

Keywords: Innovativeness of paper, Research topics;Information Science, Latent Dirichlet Allocation（LDA）, Support Vector Machine（SVM）