Multivariate statistical modeling for information integration via sparse learning

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K15947
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Statistical science
• Research Institution: The University of Electro-Communications
• Principal Investigator: KAWANO SHUICHI 電気通信大学, 大学院情報理工学研究科, 准教授 (50611448)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 機械学習 / スパース学習 / 主成分回帰 / 一般化線形モデル / オンライン学習 / 順序カテゴリカルデータ / 情報量規準 / ベイズモデリング

Outline of Final Research Achievements

We developed statistical methods for information integration that can learn a structure between data and provide a predictive model based on the given structure simultaneously. In particular, we presented a one-stage procedure for principal component regression and developed various data analysis techniques based on sparse modeling and online learning. We also proposed a procedure for selecting a value of tuning parameters in sparse modeling from the viewpoint of information theory and Bayesian approach. We applied the proposed methods into various datasets including mouse consomic strain data.

Free Research Field

統計科学