2023 Fiscal Year Final Research Report
Fast statistical analysis for large spatial datasets with temporal information
| Project/Area Number |
21K13273
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 07030:Economic statistics-related
|
|---|
| Research Institution | Kanto Gakuin University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2021-04-01 – 2024-03-31
|
| Keywords | 確率場 / 大規模空間データ / 多重解像度近似 / 時空間データ / 状態空間モデル / カルマンフィルタ / データサイエンス |
|---|
| Outline of Final Research Achievements |
In this program, I mainly conducted the following two research topics in the fast statistical analysis for large spatial or spatio-temporal datasets. First, I proposed a new fast computation method of the Kalman filter for large spatio-temporal datasets, which was referred to as a multi-resolution filter via linear projection, by using a multi-resolution approximation via linear projection which was proposed in the previous Grant-in-Aid for Early-Career Scientists. Furthermore, I extended the proposed method to nonlinear and non-Gaussian state-space models. Second, I considered a modification of the multi-resolution approximation via linear projection by using the covariance tapering, which can resolve the artificiality in the prediction surface. I evaluated the modified method through the real data analysis. I also conducted other researches related to spatial statistics.
|
| Free Research Field |
空間統計学
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究課題において主に得られた研究成果は，大規模空間データや大規模時空間データに対する新しい高速統計解析手法を提案したものであり，そのうちの一部は統計学における国際学術誌に投稿中である．特に，１つ目の研究成果である「線形射影を用いた多重解像フィルタ」は，時空間データが非ガウス性や非定常性を持つ場合にも適用可能であるだけでなく，大規模空間データが時々刻々と観測される状況でもリアルタイムで高速に統計解析を実行できる．提案手法は地価，交通量，人工衛星から得られる大気中の水蒸気量といった時空間データの統計解析に有用であり，不動産市場の分析，交通渋滞の削減，気候変動問題といった応用と結びついている．
|
