关键词: 循证数字人文, 一手证据, 共词分析, 《鲍氏国策》

Abstract: [Purpose/Significance] Evidence acquisition is one of the most critical factors affecting evidence-based digital humanities research. The first-hand evidence contained in the ancient literature works is an important way to carry out digital humanities research, and thus, the purpose of this research is to shed light on the evidence-based digital humanities research process based on the empirical analysis of Bao's Zhan Guo Ce, which is one of the most influential books in Chinese history. [Method/Process] In the face of rich and diverse Chinese ancient literature works, it is of theoretical and practical value to build an independent knowledge system with Chinese characteristics based on the evidence-based paradigm of digital humanities. For this reason, the present research used the natural language processing (NLP) method to analyze Bao's Zhan Guo Ce in Jiayan Library, which is tailored for the NLP analysis of Chinese ancient literature works. By using co-word analysis, this research comprehensively discusses how digital humanities researchers carry out systematic research based on first-hand evidence from ancient literature via word frequency analysis, visualization of co-words, cluster analysis, centrality degree analysis, etc. Social network analysis (SNA), NetworkX algorithm and co-word visualization procedure are applied to give us insight into how to extract the first-hand evidence from ancient literature works. [Results/Conclusions] The key results include a procedure on how to extract first-hand evidence from ancient literature works like Bao's Zhan Guo Ce, in digital humanities research via Python. Specifically, the procedure includes basic word frequency indicators, a tool of removal of stop words, process of recognition and removal of ambiguous words. Furthermore, this study also takes Bao's Zhan Guo Ce as an example to show the basic procedure of analyzing first-hand evidence in digital humanities research by using a series of statistical analysis methods and indicators such as co-word network visualization, clustering coefficient, centrality degree, and structural hole recognition. The procedures, tools and methods demonstrated in this study are expected to provide reference for completing the evidence-based digital humanity research paradigm of first-hand evidence. Thus, the procedures, tools, statistical indicators and algorithm demonstrated in this research are expected to provide a foundation for building an independent knowledge system of evidence-based digital humanities with Chinese characteristics.

Key words: evidence-based digital humanity, first-hand evidence, co-word analysis, Bao's Zhan