網路的發展使得資料量愈來愈龐大，使用者要從大量的資訊中找到想要的訊息是非常困難的，因此本研究利用自建的ontology架構來建立ontological user profile幫助記錄使用者興趣。本研究主要分三大部份：自動建立ontology與user profile建立，最後利用兩者結合的ontological user profile應用在文件的推薦上。在自動建立ontology部份，本研究參考A. Sieg等人(Sieg, Mobasher, & Burke, 2007)在2007年定義的名詞，將階層式概念架構稱作ontology，其中概念當作分群文件的集合。因此本研究使用動態文件階層分群來建立ontology架構，接著我們針對部份流程進行優化，改善其準確度與效能。 User profile部份，本研究利用自建的ontology架構再加上使用者接觸過的文件與行為因子來建立explicit profile與implicit profile，此user profile可以表達使用者不同時期的興趣以及達到使用者興趣的延伸，可以幫助使用者找到符合興趣的文件。 在應用上，將自建的ontology架構為文件進行階層式分類，並利用user profile讓使用者能夠加入自己讀過的文件，將此架構稱作ontological user profile，接著為使用者進行文件上的推薦，協助使用者在該領域上能夠結合自己的興趣進行融會貫通。 ;With the development of the internet, the data amount is growing in a rapid speed. Therefore it is getting more and more difficult for users to find the information from large amounts of data. As the result, this study uses self-built ontology to create ontological user profile for users recording their interests. The study consists of three parts: automatically constructing ontology, user profile creation and application of ontological user profile. The first part, we refer the noun that A. Sieg(Sieg et al., 2007) defined in 2007,we define our ontology as a hierarchy of concepts, where the concepts are utilizes for the categorization of documents. Therefore, this study constructs ontology by dynamically hierarchical clustering on documents. The second part of user profile creation, we use self-built ontology with the documents that user read and behavior factors to create explicit and implicit profile. Those user profiles express the user′s interests in the different periods, and help users to find the documents that they are interested in. The final part of the application, we use the construction of the ontology to cluster the documents, and make use of user profile for adding documents by users themselves. Through the ontological user profile, we can recommend the documents to users and help users combined their interests on the reading process.
