A Datamining Approach to Auto-extraction of Browsing Structures of Web Materials

• Main Authors: Chien-Chung Kuo, 郭建中
• Other Authors: Feng-Hsu Wang
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/68866050710419577772

碩士 === 銘傳大學 === 資訊管理研究所 === 89 === Recently, the rapid grow up of Internet stirs the development of web-based learning environments. As compared with conventional CAI systems, web-based learning environments are able to record learning behaviors of students, and hence could accumulate a large amount of learning log data. As a result, there is a need for analyzing methodologies to discover potential information from the huge log data. Such discovery of mearningful learning information might be of great help for teachers to be aware of student’s learning situations and to react properly in a responsive manner. In this paper, we propose a methodology of mining web-material browsing structures. The main concept of the methodology is to integrate the techniques of Hierarchical Clustering, Association-Rule Mining and Sequential-Patterns Mining. The mined information include the material-clusters and the sequences among the material-clusters. The mining results can be used not only to give browsing advice for learners but also enable teachers to identify some interesting or unexpectable learning patterns, based on which they can make improvements to their teaching process. Finally, this study conducts two experiments. In the first experiment, we use the artificial data to investigate the influence of some parameters on the mining results. In the second part, experiments are conducted on actual student’s log data to investigate the effectiveness of the methodology, and the results are satisfying.