| Summary: | 碩士 === 世新大學 === 資訊管理學研究所(含碩專班) === 99 === This work focuses on how to development a categorizing system that can find the appropriate classification for on-line learning resources, which must follow some document standards, automatically. The proposed system can help users to share their creativities without the dilemma of following the standards of learning resources.
There are three major modules: learning object analysis, learning resource classification, learning resource document matching, in this proposed work. "Learning Object Analysis" module can find out the information of title, keyword, and description for users' uploading resources. "Learning Resource Classification" module takes in charge of finding corresponding keywords of learning resources in the document database of the E-Learning web site. The similarity of users' uploading resources and documents in the web site document database is done by "Learning Resource Document Matching" module. The proposed automatic categorizing system for on-line learning resources can suggest the appropriate classification and standard information for users' uploading resources after performing these modules.
From the simulation results, it is obvious that "description" attribute is more important than "title" and "keyword" attributes when considering the accuracy of classification. The more training data the system uses, the higher accuracy of classification the system has. Computational time will increase a little bit when the critical attribute is used to do classification.
|
|---|
