2018 Fiscal Year Final Research Report
Theory and application of unsupervised learning for Network data and functional data
| Project/Area Number |
16K16024
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Statistical science
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Keywords | 関数データ解析 / グラフ分割 / クラスタリング / 教師なし学習 / 半教師なし学習 / 機械学習 |
|---|
| Outline of Final Research Achievements |
With recent advances in computer and measurement technologies, big and complicated data have been common in various application fields, and thus the importance of unsupervised learning has been recognized. In this research, I dealt with the following 4 research topics related to unsupervised learning for the complicated data: (1) I studied theoretical properties of graph-partitioning clustering method, (2) I developed a new semi-supervised learning method for functional data with theoretical guarantees and used the proposed algorithm to identify handball players who are at-risk for anterior cruciate ligament (ACL) injury based on ground reaction force data, (3) I developed a general approach via multiscale bootstrap to selective inference with theoretical guarantees, (4) I developed a new regularized subspace clustering algorithm for functional data which is based on a cluster-separation criterion in the finite-dimensional subspace.
|
| Free Research Field |
統計科学，機械学習
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究では，実社会への応用を想定し，応用上重要な問題に対して，新しい教師なし学習法の開発や理論的性質の解明を行っている．例えば，研究(1)では教師なし分類法において金字塔と呼べる広く用いられているクラスタリング法に関して，これまで明らかとなっていなかった重要な理論的性質を解明している．さらに，本研究では，理論研究にとどまらず，実社会の問題への応用を実際に行っている．実際に，研究(2)ではスポーツ医学の分野において，提案手法を適用することで怪我のリスクのある選手の特定に成功している．
|
