土橋 喜

With the spread of online distance learning and digital teaching materials, research analyzing the learning logs from learning management systems and uncovering knowledge that can be used to improve classes is becoming increasingly important. The interquartile method and the interquartile range were applied to the classification of learning patterns, the creation of cluster heat maps, and to outlier detection to improve the robustness of this research. Comparison of extraction methods also clarified that multiple methods were necessary to extract outliers. The analysis results of learning log classification, cluster heat mapping, and outlier detection can also be used to explain the basis of insufficient student effort when teachers make academic interventions. The experimental data showed that learners in need of academic intervention can be categorized by visualizing class engagement with a cluster heat map and outlier extraction.

This study proposes a method for a large number of university students to analyze the synchronization ratio in opening teaching materials in classes using Moodle learning management system (LMS) and online teaching materials. Furthermore, this research analyzes the results in terms of both educational technology and ergonomics and discusses the improvement in the interaction between teachers and learners. Learners’ reactions are influenced not only by the teacher's instructions, but also by various factors such as the quality of the content, the method of presenting the teaching materials, the performance of the software, and the performance of the monitor screen. Therefore, using the proposed method, the analysis was limited to the clickstream of the teaching material accumulated in the LMS. The time-series cross-section displays the synchronization ratio of teaching materials in the class in graphs and tables at intervals of between 1 and 5 minutes. Moreover, since the synchronization ratio is based on the time-series cross-section for each learner, one can observe the synchronized learners, the learners who open the teaching materials late, and the learners who never open the teaching materials.

Webから収集した文献を対象に論文サーベイを行うときの支援を目標に,文献のテキスト部分からキーワードの共起関係を抽出し,それらを用いて知識連鎖を生成して可視化し,発想活動を促進するインタラクティブな支援方法を提案する.テキストマイニングによって抽出されたキーワードから,同一センテンス上で隣接するキーワードに関連性があるものと見なし,これら2つキーワードを結合して共起関係を生成した.開発したTermLinkerシステムでは,共起関係のつながりが概念ネットワークの形で知識連鎖のように描画され,利用者は描画された共起関係とインタラクションを繰り返しながら発想活動を行うことができる.

本稿では大規模複雑な問題解決における仮説生成支援について論じている。地球環境問題を具体的な例として取り上げ、文献情報の分析を利用して、問題の関連構造を説明するための仮説生成を支援することをめざす。地球環境問題は緊急に解決が要請されている大規模複合問題であり、このような問題解決支援のひとつとして、仮説生成支援が有効であることを論じ、理論に基づくシステムの開発方法を提案し、試作を行なっている。This paper proposes a method for aiding hypothesis formation in very complex problem solving. Earth environmental problem is a typical example of very large complex problem and immediate solution is required. A key to solution of earth environmental problem is to capture the relations between problems. This paper propose how to elicit hidden relations in earth environmental problems using analysis of hyperdocuments and functions of digital library.