• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Cluster analysis for grouping statistical models and its application

Research Project

Project/Area Number 19K11862
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 60030:Statistical science-related
Research InstitutionKyushu University

Principal Investigator

Hirose Kei  九州大学, マス・フォア・インダストリ研究所, 准教授 (40609806)

Project Period (FY) 2019-04-01 – 2022-03-31
Project Status Completed (Fiscal Year 2021)
Budget Amount *help
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2021: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2020: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2019: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
Keywordsクラスタリング / クロスバリデーション / 交差検証法 / 多変量解析 / Generalized Lasso / 予測モデリング / 予測モデル / 正準判別 / 合計値予測
Outline of Research at the Start

近年,諸科学の分野で多様性を伴うビッグデータが取得されている.入力データもしくは出力データに多様性を伴う場合,単一の回帰分析や判別分析では予測精度が高くないことが多い.そこで,複数の予測モデルを構築することが考えられるが,あまり大量に予測モデルを作りすぎてもかえって予測精度が向上しないことがある.そこで本研究では,複数の予測モデルをグループ化する.これを実現するために,予測モデルに対するクラスター分析を行う.目的関数として,予測誤差に基づく関数を定義することにより,予測精度を向上させる.このモデルに含まれるパラメータを高速に推定するために,効率的な計算アルゴリズムを提案する.

Outline of Final Research Achievements

We constructed flexible statistical modeling for capturing complex structures in data. Specifically, instead of using a single regression or discriminant analysis, we constructed multiple statistical models and grouped them; then, a prediction was performed for each group. To group the statistical models, a cluster analysis was performed. Conventional cluster analysis adopted a distance matrix. On the other hand, we defined a function based on the prediction error to improve the prediction accuracy. Furthermore, an efficient computational algorithm to perform clustering was established.

Academic Significance and Societal Importance of the Research Achievements

近年、ディープラーニングを用いたデータ解析が主流となっているが、ディープラーニングは、画像やテキストなど、サンプルサイズが十分に大きい場合に精度の高い予測モデルが構築できる。一方で、遺伝子データや電力需要量データ、材料データ等、ディープラーニングが実行できるほどの多くの観測が得られないことがある。本研究では、このような場合に、できるだけ精度良く予測できる柔軟なモデルを提案した。また、モデルを新たに構築しただけでなく、高速なアルゴリズムの提案、さらにはRパッケージの公開も行った。

Report

(4 results)
  • 2021 Annual Research Report   Final Research Report ( PDF )
  • 2020 Research-status Report
  • 2019 Research-status Report
  • Research Products

    (16 results)

All 2022 2021 2020 2019

All Journal Article (9 results) (of which Peer Reviewed: 5 results,  Open Access: 8 results) Presentation (7 results) (of which Int'l Joint Research: 3 results,  Invited: 2 results)

  • [Journal Article] Interpretable Modeling for Short- and Medium-Term Electricity Demand Forecasting2021

    • Author(s)
      Hirose Kei
    • Journal Title

      Frontiers in Energy Research

      Volume: 9

    • DOI

      10.3389/fenrg.2021.724780

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Sparse multivariate regression with missing values and its application to the prediction of material properties2021

    • Author(s)
      Teramoto Keisuke、Hirose Kei
    • Journal Title

      International Journal for Numerical Methods in Engineering

      Volume: 123 Issue: 2 Pages: 530-546

    • DOI

      10.1002/nme.6867

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Relationship between gene regulation network structure and prediction accuracy in high dimensional regression2021

    • Author(s)
      Okinaga Yuichi、Kyogoku Daisuke、Kondo Satoshi、Nagano Atsushi J.、Hirose Kei
    • Journal Title

      Scientific Reports

      Volume: 11 Issue: 1

    • DOI

      10.1038/s41598-021-90791-6

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Sparse multivariate regression with missing values and its application to the prediction of material properties2021

    • Author(s)
      Keisuke Teramoto, Kei Hirose
    • Journal Title

      arXiv

      Volume: arXiv:2103.09619 Pages: 1-18

    • Related Report
      2020 Research-status Report
    • Open Access
  • [Journal Article] Event Effects Estimation on Electricity Demand Forecasting2020

    • Author(s)
      Hirose Kei、Wada Keigo、Hori Maiya、Taniguchi Rin-ichiro
    • Journal Title

      Energies

      Volume: 13 Issue: 21 Pages: 5839-5839

    • DOI

      10.3390/en13215839

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Effects of underlying gene-regulation network structure on prediction accuracy in high-dimensional regression2020

    • Author(s)
      Okinaga Yuichi、Kyogoku Daisuke、Kondo Satoshi、Nagano Atsushi J.、Hirose Kei
    • Journal Title

      bioRxiv

      Volume: なし Pages: 1-12

    • DOI

      10.1101/2020.09.11.293456

    • Related Report
      2020 Research-status Report
    • Open Access
  • [Journal Article] Interpretable modeling for short- and medium-term electricity load forecasting2020

    • Author(s)
      Kei Hirose
    • Journal Title

      arXiv

      Volume: arXiv:2006.01002 Pages: 1-29

    • Related Report
      2020 Research-status Report
    • Open Access
  • [Journal Article] L1正則化法に基づく因子分析および構造方程式モデリングの最近の展開2020

    • Author(s)
      廣瀬慧
    • Journal Title

      計算機統計学

      Volume: -

    • NAID

      130007877407

    • Related Report
      2019 Research-status Report
    • Peer Reviewed
  • [Journal Article] 因子分析モデルにおける構造正則化2020

    • Author(s)
      廣瀬慧
    • Journal Title

      京都大学 数理解析研究所 講究録

      Volume: 2133 Pages: 1-10

    • Related Report
      2019 Research-status Report
    • Open Access
  • [Presentation] Sparse multivariate regression with missing values and its application to material properties prediction2022

    • Author(s)
      Hirose, K., and Teramoto, K.
    • Organizer
      IASC-ARS2022 (The 11th Conference of the IASC-ARS)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 回帰モデルの合計値予測とクラスタリング2021

    • Author(s)
      廣瀬 慧、増田 弘毅
    • Organizer
      2021年度 統計関連学会連合大会
    • Related Report
      2021 Annual Research Report
  • [Presentation] クロスバリデーションに基づく正準判別のクラスタリングとその高速化2021

    • Author(s)
      三浦 完太, 廣瀬 慧
    • Organizer
      2021年度 統計関連学会連合大会
    • Related Report
      2021 Annual Research Report
  • [Presentation] Hierarchical multiclass discriminant analysis via cross-validation2021

    • Author(s)
      Hirose, K., and Miura, K.
    • Organizer
      The 4th International Conference on Econometrics and Statistics (EcoSta 2021)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 電力需要予測のための統計モデルとソフトウェア2020

    • Author(s)
      廣瀬慧
    • Organizer
      2020年度 統計関連学会連合大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 電力需要の短期予測のための統計モデリング2019

    • Author(s)
      廣瀬 慧,増田 弘毅
    • Organizer
      2019年度統計関連学会連合大会
    • Related Report
      2019 Research-status Report
  • [Presentation] Cluster-based multiclass linear discriminant analysis2019

    • Author(s)
      K. Hirose, K. Miura, A. Koie
    • Organizer
      The 3rd International Conference on Econometrics and Statistics (EcoSta 2019)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research

URL: 

Published: 2019-04-18   Modified: 2023-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi