共词分析预测研究前沿的表面效度研究:基于自然语言处理

The Face Validity of Research Frontier by Co-words Analysis Forecast Based on the Natural Language Process

本研究使用自然语言处理方法,以动物学领域取近30年发表于中文学术期刊并收入CNKI数据库的57 438篇学术文献为分析对象,基于共词分析对研究前沿预测的表面效度进行检验。研究发现,基于全文而探测到的研究前沿表面效度最高,而基于题名和关键词而探测到的研究前沿则表面效度很低。此外,基于摘要而探测到的研究前沿的表面效度具有不稳定性,因此摘要在研究前沿探测中的价值存疑。本研究初步回答了基于共词分析的科学计量工具与指标在研究前沿探测方面是否具有及在多大程度上具有有效性的问题。

The present study conducted a natural language process to test the face validity of co-words analysis on research frontier forecast based on the academic papers in Zoology research community published within recent 3 decades and included in CNKI data- base. Findings of this research included that the research frontier forecast based on full texts had the highest face validity comparing to the titles and keywords. However, the explored face validity of research frontier based on abstracts was unstable. The evidence of this research initially provided the answer to the question that whether or not the research frontier forecast based on the co-words analysis had validity or on what extent of the validity existed.