數位人文視野下的歷史分期法：以《人民日報》（1946—2016）中深圳主題報導為例

A Digital Humanities Research on the Method for Periodization of History: The Example Based on the Shenzhen-related News Reports in the People's Daily (1946-2016)

本文研究目的是希望能從數位人文研究的角度，提出一種質量並重的歷史分期法，以有別於過去多從個人主觀視角出發進行定性的歷史分期工作。本文以1946—2016年間《人民日報》中所有標題中包含「深圳」一詞的報導文章為研究範圍，共有3,203篇，運用（1）人工主觀分期；（2）數值高峰分期；（3）關鍵詞叢異同分期等三種方法，對為期70年的涉深報導進行分期工作。在研究過程中發現，上述三種分期法雖都可進行歷史分期工作，然皆有其侷限性，即這些分期法都含有個人主觀性，在進行分期工作時往往因人而異。因此，本文提出第四種歷史分期法，嘗試以量化數據作為客觀分期之標準，為長時段的歷史研究提供一分期工作上的新方法與新視野，此即為本文研究結果與價值所在。

This study aims to provide a method of historical periodization based on digital humanities that could better combine qualitative and quantitative analysis than previous ones. This study singled out 3,203 news reports in the People's Daily, 1946 to 2016, whose headlines contained "Shenzhen" and respectively applied 1) artificial periodization, 2) peak number periodization, and 3) keyword cluster periodization to these Shenzhen-related news reports. The results show that three methods of periodization respectively have their limitations on objectivity. This paper thus suggests an automatic periodization based on quantitative datasets and this data-driven method could provide a novel perspective for the long-term historical analysis in the future.