Sparse estimation via regularization and its application to ultra high-dimensional data
Project/Area Number |
24700277
|
Research Category |
Grant-in-Aid for Young Scientists (B)
|
Allocation Type | Multi-year Fund |
Research Field |
Statistical science
|
Research Institution | Osaka University |
Principal Investigator |
HIROSE Kei 大阪大学, 基礎工学研究科, 助教 (40609806)
|
Research Collaborator |
YAMAMOTO Michio 京都大学, 医学(系)研究科(研究院), 助教 (50721396)
|
Project Period (FY) |
2012-04-01 – 2015-03-31
|
Project Status |
Completed (Fiscal Year 2014)
|
Budget Amount *help |
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2014: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2013: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2012: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
|
Keywords | スパース推定 / L1正則化 / モデル選択 / 因子分析 / 因子分析モデル / 因子回帰モデル / 非凸ペナルティ / lasso / 線形回帰モデル |
Outline of Final Research Achievements |
Sparse estimation via L1 regularization enables us to analyze the ultra high-dimeniosnal data, and it has been rapidly developed in the recent years. I made two research achievements in this project. First, I developed an efficient algorithm that estimates the tuning parameter in sparse regression modeling. Second, I proposed the L1 regularization procedure in the factor analysis model, and investigated the relationship between the traditional rotation technique and regularization procedure. I showed that the regularization is viewed as the generalization of the rotation method. I have also made a free software package in R (msgps and fanc).
|
Report
(4 results)
Research Products
(20 results)
-
[Journal Article] Full information maximum likelihood estimation in factor analysis with a lot of missing values2015
Author(s)
Hirose, K., Kim, S., Kano, Y., Imada, M., Yoshida, M. and Matsuo, M.
-
Journal Title
Journal of Statistical Computation and Simulation
Volume: 印刷中
Issue: 1
Pages: 91-104
DOI
Related Report
Peer Reviewed
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-