許玉敏 Yu-Min Hsu, 鄭文信 Wen-Hsin Cheng

以知識本體為基礎建置澎湖族譜語意資料庫
The Genealogy Database of Penghu County Based on Ontology

家譜資料對澎湖學來說具有相當的重要性，尤其是在日據時代之前，關於澎湖的文獻資料較為稀少。家譜是少數以紙本形式保留下來的澎湖地方文獻資料，其中記載了各家族的組織、繁衍、發展、遷移、婚姻、教育……等資訊。然而家譜的數量固然眾多，其資料類型大多只保留每一個人物片斷式的外部特徵描述，每一筆資料單獨來看通常不見其意義。只有將所有資料匯整成一個巨大的知識網絡，才能夠突顯這些文獻所隱含的意涵。然而，以人工方式來進行統計分析的工作，往往有如披沙揀金，事倍功半。本研究希望能夠藉由資料庫的建置來提供一個可能的研究途徑。
澎湖族譜資料庫將以知識本體為基礎，並利用美國史丹福大學開發的 protégé 軟體建置知識本體庫、開發資料庫。其中有兩個重要知識範疇：第一個是家庭關係的知識本體架構，這個部分將參考國際族譜規範GEDCOM的格式。GEDCOM最初於1985年開發，由耶穌基督後期聖徒教會家族史部門擁有和管理，專門用來保存和交換族譜訊息的數據格式。它具有體積小、便於流傳與導入不同家譜式程式使用的優點，也是目前全球最通用的族譜存儲格式。然而，因家庭、婚姻文化的不同，GEDCOM規範並無法滿足華人社會的家庭狀況；尤其面對華人傳統家庭一妻多妾的家庭結構，有些親屬關係需要另外再做定義。以《呂氏家乘》為例，此家譜包含了一些較為複雜的身份或關係。例如：如兼祧且所兼兩房均有嫡妻、過繼、過繼後又回到本家、夫死後寡婦招贅、妾生子被記入族譜、入贅與其他房等等關係，這些都是建立家族關係的知識本體需要處理的條目。第二個重點是地名的知識範疇。有許多家譜會記載祖籍及成員的遷移狀況。還有某些地區之間的來往，如嫁娶、收養、工作、遷移等等。此外，澎湖地區的地名亦常有一地多名的現象。再者，有些地名的範圍或行政區域的劃分也因時而異，將進行這些問題的討論與開發解決方案。為了資料的運算，以及將來能夠運用GIS來完成視覺化的呈現，地名的知識本體也是必須完成的工作。
本研究嘗試以知識本體為基礎，開發互動式家譜輸入平台。目的是為了能夠更有效率的輸入、彙整澎湖地區大量的家譜文獻。
此平台的輸入介面如圖2
族譜的完成圖如圖3
在網頁上（圖1），將游標移到任一人物上面，就可以進行新増相關人物，或是修改、刪除其資料的工作：
平台所輸入的族譜資料能夠轉成GEDCOM格式輸出，有利於將來與其他資料，甚至是全球的資料接軌。
為了將來有機會藉此平台推廣族譜文化，讓一般大眾都能加入尋根的行列，此平台以開發直覺式的、互動式的輸入介面為目標，普通民眾也能輕易學會輸入自家族譜。是推廣家族文化、地域文化的有利工具。將來可以運用此平台，鼓勵澎湖居民或後裔建立、查詢、分析自己的族譜，尋根溯源，更瞭解自己的傳承根基。

The purpose of this research is to develop an ontology Database that allowed researchers to input, calculate, and analyze a huge number of genealogical documents in the Penghu area efficiently.
The genealogy has significance for studying Penghu history since there are only a few other documents about Penghu history. The genealogy is one of the few local documents which is remaining. It contains information about family trees, reproduction, development, migration, marriage, education, etc. Although there are many genealogies in Penghu, there is very little information about each character in it. Therefore, it doesn't make much sense to study each character individually. These data are meaningful only if the research is on a group basis. Building an ontology Database platform can help us achieve this goal without manual statistical calculations that usually consume numerous time and energy.
The database is based on ontology. This study uses the protégé software developed by Stanford University to build the ontology database. There are two works we need to finish before we developed the database. First, developing the ontology of family kinship. The ontology of family kinship in this study is based on the GEDCOM Standard but adjust the structure to fit the Chinese family structure.
Second, developing the ontology of place names. There are many records involved the interaction of people from multiple places, such as marriage, work, adoption, immigration, and so on. It is not rare that two places share the same name, or one place has multiple names in Penghu. In addition, some places have different names at different times. It is necessary to build the ontology of place names to avoid misunderstanding and to have visualized graphs with GIS tools.
An interactive genealogy platform helps the development of the database efficiently. Here is the input interface of the platform(fig.2)
Here is an example that shows a family tree of a family(fig.3)
This picture shows how to add, delete, modify an individual(fig.1)
Any family tree that the platform collected can transform to the GEDCOM format which allowed researchers to export the file. An interactive genealogy platform is a useful tool that helps popularize genealogy. This platform can be used to encourage Penghu residents or descendants to establish, inquire, and analyze their genealogy, trace the roots, and understand their family history better.