| Project/Area Number |
24K22613
|
|---|
| Research Category |
Grant-in-Aid for Research Activity Start-up
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
0107:Economics, business administration, and related fields
|
|---|
| Research Institution | Akita International University |
|---|
Principal Investigator |
Yanchenko Eric 国際教養大学, 国際教養学部, 助教 (61000146)
|
|---|
| Project Period (FY) |
2024-07-31 – 2026-03-31
|
| Project Status |
Granted (Fiscal Year 2024)
|
| Budget Amount *help |
¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2025: ¥520,000 (Direct Cost: ¥400,000、Indirect Cost: ¥120,000)
Fiscal Year 2024: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
|
|---|
| Keywords | Bayesian inference / High-dimensional data / Shrinkage priors / Big data / Bayesian statistics |
|---|
| Outline of Research at the Start |
The goal of this project is to develop a novel shrinkage prior model for the sparse grouped regression setting by leveraging the R2D2 prior framework.
|
| Outline of Annual Research Achievements |
We completed the paper "The Group R2D2 Shrinkage Prior for Sparse Linear Models with Grouped Covariates," extending the R2D2 prior framework to the group regression setting. This work is currently under review for publication.
The grant has also aided two other projects. The work was extended to the ordinal regression setting in "Pseudo-R2D2 prior for high-dimensional ordinal regression," currently in major revision at Statistics and Computing. Also, "BOPIM: Bayesian Optimization for Influence Maximization in Temporal Networks" has been accepted subject to minor revisions at Technometrics.
|
| Current Status of Research Progress |
Current Status of Research Progress
1: Research has progressed more than it was originally planned.
Reason
We completed the initial goal of this grant to develop the R2D2 prior for the group regression setting. Additionally, the grant has aided me in completing two other projects (mentioned above). I was also invited to present some of this work at Kobe University in December 2024. For these reasons, I believe things are progressing better than expected. Please note that due to the relatively slow timeline of reviews at statistics journals, these papers have not been published yet, but we anticipate them to be in the coming year.
|
| Strategy for Future Research Activity |
There are several directions I am excited to pursue this year. First, I was invited to present some of the research accomplishments at a Bayesian workshop in Tokyo this August 2025. Additionally, I continue to work with my collaborators on future extensions of our methods. For example, we are currently studying the differences between enforcing sparsity in the mean function vs. the covariance function in Gaussian Process models. These ideas are closely related to the shrinkage priors used in the original grant proposal.
|
