| Summary: | 博士 === 國立成功大學 === 工程科學系碩博士班 === 97 === The energetic development of the Internet, especially on the web page interaction technology, has made distant e-learning systems become more and more realistic and popular in the past ten years. Problems due to technology shortcomings, however, gradually emerge when using current e-learning systems, among which, how to compose personalized courses dynamically in accordance with a user’s intention is the largest challenge. One of the advantageous prospects of e-learning systems comes from the easiness of achieving personalized learning, which is virtually impossible in traditional classrooms; but the lack of proper technologies has been blocking the dream from coming true. This thesis proposed a semantic-aware approach that makes an e-learning system able to infer a user’s query and then recommend a personalized course according to the user’s preference and intention. The proposed approach uses a semi-automatic ontology constructing mechanism, developed in this research, to build domain knowledge ontology of different courses. By using a constructed ontology, the approach can analyze a user’s query and understand what concepts in a specific domain the user is intending to learn. It then uses a hybrid recommendation model, developed also for the proposed approach, to recommend suitable learning objects according to a user’s preference and intention. In the last phase of the approach, the personalized course composition, an adapted discrete particle swarm optimization is used to promote the performance of picking suitable learning objects and a smooth reading order is constructed for the user’s comfortable learning and reading. From the experimental results, it shows that personalized courses dynamically composed by the proposed approach can satisfy different users’ needs with their feedbacks indicate that the recommended domain concepts conformed to their learning intentions and the picked learning objects fit their preferences.
|
|---|
