Development of SVM-Based IR System for Professional Development at Japanese Universities
| Project/Area Number |
26282057
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Partial Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Educational technology
|
|---|
| Research Institution | Tokyo Metropolitan University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
立石 慎治 国立教育政策研究所, 生徒指導・進路指導研究センター, 研究員 (00598534)
大森 不二雄 首都大学東京, 大学教育センター, 教授 (10363540)
永井 正洋 首都大学東京, 大学教育センター, 教授 (40387478)
林 祐司 首都大学東京, 大学教育センター, 准教授 (40464523)
椿本 弥生 公立はこだて未来大学, システム情報科学部, 准教授 (40508397)
松河 秀哉 大阪大学, 全学教育推進機構, 助教 (50379111)
渡辺 雄貴 東京工業大学, 教育革新センター, 准教授 (50570090)
松田 岳士 首都大学東京, 大学教育センター, 教授 (90406835)
|
|---|
| Co-Investigator(Renkei-kenkyūsha) |
TAKAMORI Tomotsugu 福島大学, 総合教育研究センター高等教育開発部門 (80583103)
|
|---|
| Research Collaborator |
YANAGIURA Takeshi Postsecondary Analytics, 代表
|
|---|
| Project Period (FY) |
2014-04-01 – 2017-03-31
|
| Project Status |
Completed (Fiscal Year 2016)
|
| Budget Amount *help |
¥9,620,000 (Direct Cost: ¥7,400,000、Indirect Cost: ¥2,220,000)
Fiscal Year 2016: ¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000)
Fiscal Year 2015: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2014: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
|
|---|
| Keywords | 教学ＩＲ / ＯＲ / 機械学習 / ラーニングアナリティクス / ＦＤ・ＳＤ / SVM / 教授学習支援システム / FD・ID / IR / OR / ＩＲ / 機関分析 / データマネージメント |
|---|
| Outline of Final Research Achievements |
In this study, the IR system that is useful for university faculty and staff in supporting students has been investigated. We developed an early alert system, or ‘Risk Detector’ (hereafter RD), to identify at-risk students before their potential risks become aware and to prevent the students from dropouts and holdovers. RD was expected to pinpoint high risk students with the machine learning method called Soft Margin Support Vector Machine and to indicate a list of metrics of each student as well as the prediction of dropout.
After feeding actual students’ data into the RD, its accuracy rate of prediction was 93%, which meant it provided highly reliable results of discriminant analysis and reasonable precision. In addition, RD was recognized trustworthy by faculty members who assessed it. On the contrary, poor UI and intelligibility of each value on RD were pointed out and discussed as future issues.
|
Report
(4 results)
Research Products
(17 results)
