Development of statistical modeling with regularization for large scale data

2015 Fiscal Year Final Research Report

• Project/Area Number: 25730017
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Statistical science
• Research Institution: Kyushu University
• Principal Investigator: Matsui Hidetoshi 九州大学, 数理(科)学研究科(研究院), 助教 (90633305)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 関数データ解析 / スパース正則化 / モデル選択

Outline of Final Research Achievements

I have developed some statistical modeling strategies for repeated measurement or longitudinal data in order to extract important information from data. In this work I focused on two topics; functional data analysis and sparse regularization. The basic idea behind functional data analysis is to express longitudinal data in the form of a function and then draw information from a set of functional data. I applied sparse regularization techniques to several statistical models for functional data in order to estimate model and select functional variables simultaneously. I also derived algorithms for estimating parameters and model selection criteria for evaluating the estimated models. The proposed methods were applied to the analyses of real data sets and then examined the effectiveness of the methods.

Free Research Field

統計科学