2011 Fiscal Year Final Research Report
Solving a start-up problem of recommender systems by employing transfer learning
Project/Area Number |
21500154
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Single-year Grants |
Section | 一般 |
Research Field |
Intelligent informatics
|
Research Institution | National Institute of Advanced Industrial Science and Technology |
Principal Investigator |
KAMISHIMA Toshihiro 独立行政法人産業技術総合研究所, ヒューマンライフテクノロジー研究部門, 主任研究員 (50356820)
|
Co-Investigator(Kenkyū-buntansha) |
SHOTARO Akaho 独立行政法人産業技術総合研究所, ヒューマンライフテクノロジー研究部門, 研究グループ長 (40356340)
|
Project Period (FY) |
2009 – 2011
|
Keywords | 知識発見とデータマイニング / 推薦システム / 転移学習 |
Research Abstract |
In this project, we planed to address a start-up problem, which means the poor prediction accuracy of recommendation systems for new users or new items. For this purpose, we developed a TrBagg algorithm, which utilizes information of other users' data. We tested this algorithm on a tag recommendation task and succeeded to improve the prediction accuracy. We also tried to settle other problems related to recommender systems.
|