Estimation of Learning Activity and Learning Performance of Junior High School Student
| Project/Area Number |
18K18656
|
|---|
| Research Category |
Grant-in-Aid for Challenging Research (Exploratory)
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Medium-sized Section 9:Education and related fields
|
|---|
| Research Institution | Kyushu University |
|---|
Principal Investigator |
MINE Tsunenori 九州大学, システム情報科学研究院, 准教授 (30243851)
|
|---|
| Project Period (FY) |
2018-06-29 – 2022-03-31
|
| Project Status |
Completed (Fiscal Year 2021)
|
| Budget Amount *help |
¥5,720,000 (Direct Cost: ¥4,400,000、Indirect Cost: ¥1,320,000)
Fiscal Year 2019: ¥3,120,000 (Direct Cost: ¥2,400,000、Indirect Cost: ¥720,000)
Fiscal Year 2018: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
|
|---|
| Keywords | 振り返り文 / テキストマイニング / 機械学習 / 成績推定 / アドバイス支援 / 特徴抽出 / 中学生 / 学習状況推定 / フィードバック文自動生成 / 特徴分析 / 自動フィードバック / 目的語 / 繰り返し表現 / 深層学習 / 属性選択 |
|---|
| Outline of Final Research Achievements |
This study aims to create a system for automatically extracting characteristic expressions related to student learning behavior and learning status from their written reflections, and to utilize the extracted expressions to generate advice for improving the student learning behavior and learning ability. Specifically, we constructed a model that judges the top and bottom grades in regular examinations at a student's junior high school using the student written reflections provided by a cram school. We extracted features such as object, verb, negative, and polite expressions from the reflective sentences, weighted them, and combined them with various machine learning tools to improve the classification accuracy by up to 30% or more compared to the case where no weighting was applied.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究は，中学生の振り返り文を分析し，成績推定を行うことが可能となることを示した初めての研究である．研究の結果，振り返り文には成績推定に有用な特徴があること，各特徴と成績との間に異なる関連の強さがあること，この強さを考慮した重み付けを利用する場合と利用しない場合とでは，成績推定精度に最大30%以上の差が生じること，時期によって成績推定精度が変化し，最高の推定精度を得る機械学習器も一定しない（時期によって変化する，つまりデータの特徴が時期によって異なる）ことなどを確認した．これらから学習状況や学習行動の解釈，学習特性改善のためのアドバイス生成への応用も期待される．
|
Report
(5 results)
Research Products
(25 results)
