數位人文與時空資訊整合分析－個人歷史文本分析工具及其應用於明代倭寇研究之案例

將自然語言處理技術運用於數位化歷史文本之分析，已成為數位人文研究的一項重要課題。考量歷史事件研究之實際需求，本研究設計一套可供個人使用之歷史文本事件分析工具，並嘗試以明代持續侵擾中國沿海地區的倭寇為應用主題，以編年體之《明實錄》為實驗文本，整合歷史地名、人物與時間資料庫，對倭寇相關事例類型進行辨識、擷取，建立關聯歷史事件、人物、地名等內容之資料結構。本研究實作分析《明實錄》文本記載明代倭寇事例所涵蓋的面向，包括倭寇侵擾、官軍剿寇之戰爭事例，以及船艦往來、海防相關事務等，並藉由檢視高頻人名、地名之時間分布，對照相關文本段落進行解讀，進而運用地理資訊系統來探討明代倭寇事件及明朝海防布局之時空特性。期藉由歷史文本事件分析工具之開發與實例應用，推展數位人文之時空資訊整合分析模式於文史研究的可能性。

關鍵字

個人化歷史文本分析工具；事件分群；文本自動標註；自然語言處理；地理資訊系統

並列摘要

Applying natural language processing techniques to the analysis of digital historical texts has become an important topic in digital human studies. Considering the actual needs of historical event research, this study starts with the needs of individual researchers and designs a set of tools for historical text event analysis, and the effectiveness of the method is further evaluated by applying it to study pirate events in the Ming Dynasty. Since pirates continued to invade China's coastal areas almost throughout the Ming Dynasty, the Maritime Prohibition Policy and coastal defense arrangements in the Ming Dynasty closely related to the growth and decline of pirates' scourge. Based on some research papers and historical document text on the theme of Chinese coastal pirates, we identify and derive event groups through the individual-use text-based historical event analysis tool, and establish a data structure that links historical events, people and place names. By analyzing the type of each pirate event group, we found that it includes war cases and coastal defense affairs, and we can further track their spatiotemporal transition characteristics by name entities in each event group. This study aims to develop an individual-use text-based historical event analysis tool, and integrate geographic information system with digital humanities applications, to facilitate the possibility of digital humanities methods to historical research.