日治時期繪葉書系統資料層面連結之研究

Abstract

本研究以國立臺灣大學圖書館典藏之日治時期繪葉書為研究對象，探討數位典藏系統於現有詮釋資料下，應用資料的交互協作，提供更佳的資料呈現模式與使用體驗。研究流程上，首先，以文物編目指引（Cataloging Cultural Objects, CCO）及圖像分析等為基準，使用中央研究院詞庫小組開發之中文斷詞標註程式CkipTagger，擷取出詮釋資料欄位著錄之人、事、時、地、物及圖像、歷史等層面詞彙。接著，以量化方法比較：「內容描述ˍ摘要」欄位及「內容描述_既有文字」兩欄位於各層面連結數量之差異，藉以探討數位典藏物件於文物層次上的概念類型分布。最後，實作層面詞彙再組合，藉以探索於系統上呈現文物的應用情境。研究結果發現，「內容描述ˍ摘要」欄位之層面詞彙連結數量大於「內容描述_既有文字」，分別為12.52、1.43，且欄位間的層面詞彙連結在不同種類上具特色差異。最末，本研究基於人事時地物，提出二種敘事模式，應用量化研究證實之詮釋資料重要欄位的詞彙可用性，以案例實作詞彙層面再組合以擴充數位物件的敘事內容與脈絡，並討論前述流程之可行性與挑戰，作為未來系統資料編寫、利用、展示之參考。

This study takes postcards which were produced during the Japanese colonial period in the digital archive of the National Taiwan University Library as the research object, and explores how to comprehend the data and how the digital archive system can use the data to provide a better presentation mode. In terms of the research process, first of all, based on the Cataloging Cultural Objects (CCO) and image analysis, this study categorized results of word segmentation from CkipTagger into five facets. Next, a quantitative method was used to compare the difference in the linkage of data level between the “Description_abstract” and the “Description_text” which are the fields of metadata format. Finally, the facets are recombined to explore application scenarios on digital archive systems. This study found that the number of facet links of “Description_abstract” is greater than that of “Description_text”. The means of facet links are 12.52 and 1.43, respectively, and the distribution of facet links shows different and unique patterns. In addition to quantitative analysis, this study proposes two narrative models based on facets of people, event, time, location and object. Applying the important and useful metadata that has been confirmed by quantitative method to recombine different facets, we demonstrated the ways to enrich narrative contents and contexts for digital archives via various cases, and discussed the feasibility and challenges of the aforementioned process which could be a reference for future system data cataloging, utilization, and demonstration.