Developing distribution theory for discretely observed random field and its application to spatial statistics

2015 Fiscal Year Final Research Report

• Project/Area Number: 23500353
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Statistical science
• Research Institution: Kyushu University
• Principal Investigator: Ninomiya Yoshiyuki 九州大学, マス・フォア・インダストリ研究所, 准教授 (50343330)
• Project Period (FY): 2011-04-28 – 2016-03-31
• Keywords: 確率場理論 / ゲノム科学 / スパース推定 / 多重検定 / 統計的漸近理論 / モデル選択

Outline of Final Research Achievements

The main results are ``Development of a new multiplicity adjustment using spurious correlations'' and ``Derivation of AIC for regularization parameter selection in the LASSO.'' In the former, a new method is developed to improve the power of multiple tests, and the method is shown to be efficient in real data analysis. In the latter, for the LASSO in the framework of generalized linear models, an information criterion is derived as a unique criterion which has the same root as the classical criteria.

Free Research Field

数理統計学