Discovery scientific approach to the digital portfolio by integration of different format history data

摘要

Portfolio Assessment is the new assessment method and its effective practical use is expected particularly in Comprehensive Learning. Purpose of this study is to provide the effective tools for Portfolio Assessment. Concretely, we propose a new handling method of learning history data, named Digital Portfolio. Digital Portfolio is defined the integrated information of learning history data sets through different learning tools or environments, and extracted rules or patterns on learning activities from learning history data sets. From Digital Portfolio, information for students for self-assessment and for teachers feedback information about students' learning activities are provided. In this paper, firstly definition and construction factor & method of Digital Portfolio are described, and secondly, data structure of history data-base and condensing method of data-base using decision tree learning is explained. Particularly in this process, extracting method of association rules for multi-valued attribute data-base and continuous data-base is proposed. Finally, effectiveness of proposed method for Digital portfolio will be shown through implementation of Digital Portfolio into the supporting system for group learning Group.