面向学术资源的TSD与TDC测度及分析研究

The project prepares to introduce two new concepts named as TSD(Term Space Density) and TDS(Term Discriminative Capacity), to detect the overall characteristics status and individual distribution laws of term space and spaces of other academic objects in data collection of academic resources, to analyze the high-quality objects and the causes they formed, so that to provide effective knowledge services. The research of this project will focus on the following aspects: systematically defining the intension and extension of TSD and TDC, according to which realizing their measure algorithms and conducting validation; comprehensively detecting the significant differences of TDC-values of terms from distinct sources, and analyzing the reasons why the difference are existing, so that to evaluate the quality of indexing terms from distinct fields and with different subject characteristics, and then making constructive suggestion to improve the effectiveness of the index system; analyzing in-depth the reasons why terms belonging to academic objects spread out or clustered together in the space, and exploring the internal relationship between “good” terms and the quality of subject content and discipline innovation, so that to put forward feasibility proposals for reasonable adjustments to domain areas of academic objects studying; summarizing the characteristics and laws of spatial distribution of academic objects within a discipline, and investigating the relationship between high-quality objects and their research fields, development directions and contributions to the discipline and so on, so that to promote the healthy development of academic objects. Finishing the research report named as “Studies on measurement and analysis of Term Space Density and Term Discriminative capacity oriented academic resources” and achieving a platform for calculation and visual display of TSD and TDC will be considered as final objectives of this project.

本项目拟引入两个新的概念TSD（术语空间密度）和TDC（术语区别能力），来探测学术资源数据集合中术语及其他学术对象空间的总体特征状况和个体分布规律，分析优质对象及其形成原因，以提供有效的知识服务。研究重点：对TSD和TDC的概念进行系统定义，据此实现其测度算法并进行有效性检验；全面探测不同来源术语TDC的显著差别，分析差别存在的原因，从而对不同来源术语作为索引词的质量进行评价，并提出改善索引系统有效性的建设性意见；深入分析学术对象术语发散或聚集的原因，探讨“好”术语与内容质量、学科创新之间的内在联系，进而为学术对象研究内容的合理调整提出可行性建议；总结学科内学术对象空间分布的特征和规律，探讨优质对象与其研究内容、发展方向以及对学科的贡献等之间的关系，从而促进学术对象的健康发展。最终目标：撰写研究报告《面向学术资源的TSD与TDC测度及分析研究》；实现TSD与TDC的测算和可视化展示平台。

本课题引入了TSD（术语空间密度）和TDC（术语区别能力）两个新的指标，探讨了学术资源数据集合中术语及其他学术对象空间的总体特征状况和个体分布规律，提出并论证了测度学术数据库索引词质量的新模式，发展并完善了基于内容的学术对象科学评价的方法体系。课题探讨主要聚焦在4个方面，①TSD和TDC的测度和验证。提出了4种用于测度TDC的算法，并从微观、中观和宏观等3个层次对各种结果进行了比较分析，最终确定了TDC的测度方法：以术语╳术语矩阵作为TS（术语空间），那么在TS中移除术语前后TSD的变化量即为该术语的TDC值。②基于TSD和TDC的索引词质量分析和评价。计算来自不同子数据库和不同学科学术论文的TDC，进而在ANOVA Testing作用下探测出子数据库、学科、字段等变量会显著影响TDC，即术语质量。总体来说，学术论文用词质量普遍较高，摘要和题名字段的术语质量要高于关键词，WOS中，SCI的术语质量较高，SSCI最低，CSSCI中，考古学的术语质量最高，而教育学和马克思主义等学科的术语质量排名较低。上述研究及其结论在一定程度上弥补了基于内容特征评价术语质量的不足。③基于TSD和TDC的学科及相关对象内容分析和评价。利用TDC模型可计算出学者和期刊所使用术语的TDC值，进而根据具体术语的TDC值、TDC值的类别分布以及ANOVA Testing等分析相关学者和期刊在术语使用上的差别性。结果发现，著名学者和期刊所使用术语的质量普遍较高，同一学科的著名学者在TDC上存在显著差异，情报学CSSCI期刊主题包容性较强，而相反图书馆学期刊主题多集中于某个边缘领域，个性化较强。TDC成为了评价学术对象的又一新的指标。④基于空间密度和区别能力的学术对象的特征描述。将TDC泛化为AODC（学术对象区分能力），进而在其基础上演绎出能够在一定程度上展现期刊个性化程度和创新性力度的JDC（期刊区分能力）。在CNKI中，人文社科类期刊的JDC要明显高于理工科，档案学期刊的离散度要明显高于其它图书情报类期刊，此外JDC在时间维度上也呈现出一定的规律性。