Making Course Recommender Systems Interpretable: A Feature-aware Deep Learning-based Approach

• Item type: National Convention(1)
• 公開日: 2024-03-01
• タイトル: Making Course Recommender Systems Interpretable: A Feature-aware Deep Learning-based Approach
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: データとウェブ
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_5794
• 資源タイプ: conference paper
• 著者所属: 九大
• 著者所属: 九大
• 著者所属: 九大
• 著者所属: 九大
• 著者名: 楊, 添元; 任, 宝峰; 馬, 博軒; 木實, 新一
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: Course recommender systems can help students identify the courses that are suitable or interesting to them. Existing course recommender systems prioritize accuracy, neglecting the crucial dimension of whether students trust the system's recommendations. To address this limitation, we posit that furnishing explanations to students can enhance their trust in the system. In this paper, we introduce a novel deep learning-based course recommendation model founded on a knowledge graph, which supports path visualization and empowers students to comprehend the rationales behind the recommendations. Specifically, we replace the hidden layer in the neural network with course features, which are subsequently trained to capture students' preferences for distinct features, informing the final recommendation based on these learned weightings.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AN00349328
• 書誌情報: 第86回全国大会講演論文集; 巻 2024, 号 1, p. 399-400, 発行日 2024-03-01
• 出版者 / 言語: ja
• 出版者: 情報処理学会