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

任意の分布に従うデータに自己適応可能な学習分類子システム

Research Project

Project/Area Number 17J03593
Research Category

Grant-in-Aid for JSPS Fellows

Allocation TypeSingle-year Grants
Section国内
Research Field Intelligent informatics
Research InstitutionThe University of Electro-Communications

Principal Investigator

辰巳 嵩豊  電気通信大学, 情報理工学研究科, 特別研究員(DC1)

Project Period (FY) 2017-04-26 – 2020-03-31
Project Status Completed (Fiscal Year 2019)
Budget Amount *help
¥2,700,000 (Direct Cost: ¥2,700,000)
Fiscal Year 2019: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2018: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2017: ¥900,000 (Direct Cost: ¥900,000)
Keywords学習分類子システム / データマイニング / 雑音 / 分類問題
Outline of Annual Research Achievements

本研究では機械学習の一手法の学習分類子システムを用いて、実環境から取得されたデータから人間にとって解釈性の高いルールを獲得するために(1) 環境からの入力, 環境への出力及びその出力に対する評価である報酬が「特定」の分布に従う環境への適応,(2)入力, 出力及び報酬が「不特定」の分布に従う環境への適応, (3)学習に用いるデータ数が少なく入力の種類に偏りがあり十分に環境全体を網羅できない環境への適応に取り組む。
その目的達成に向けて本年度は(2)及び(3)に対して取り組んだ。(2)に関しては報酬に対して様々な形状の雑音が付加された環境においてこれまでに提案した学習分類子システムを適用し、雑音の形状やその強度に関する情報を与えずに獲得を目指すルールを獲得できるかを確認し、数理的な解析による手法の適用可能な雑音の限界点を明らかにした。 (3)に関してはこれまでに提案した学習分類子システムを複数の分類問題に適用し、提案手法が分類器を獲得するに当たり障害となる問題を明らかにした。
本研究の成果の社会に向けた発信として、成果を国際学会のIEEE Congress on Evolutionary Computation 2019及びThe Genetic and Evolutionary Computation Conference (GECCO) 2019のWorkshopと国内学会の進化計算シンポジウムにおいてそれぞれ1件ずつ発表した。

Research Progress Status

令和元年度が最終年度であるため、記入しない。

Strategy for Future Research Activity

令和元年度が最終年度であるため、記入しない。

Report

(3 results)
  • 2019 Annual Research Report
  • 2018 Annual Research Report
  • 2017 Annual Research Report
  • Research Products

    (14 results)

All 2019 2018 2017

All Journal Article (4 results) (of which Peer Reviewed: 4 results,  Open Access: 3 results) Presentation (10 results) (of which Int'l Joint Research: 6 results)

  • [Journal Article] Acquiring Classifiers for Bipolarized Reward by XCS in a Continuous Reward Environment2019

    • Author(s)
      Takato Tatsumi, Keiki Takadama
    • Journal Title

      ICE Journal of Control, Measurement, and System Integration

      Volume: 印刷中

    • Related Report
      2018 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Learning Classifier System Based on Mean of Reward2017

    • Author(s)
      Tatsumi, T., Sato, H., and Takadama, K.
    • Journal Title

      Journal of Advanced Computational Intelligence and Intelligent Informatics

      Volume: 21 Issue: 5 Pages: 895-906

    • DOI

      10.20965/jaciii.2017.p0895

    • NAID

      130007520198

    • ISSN
      1343-0130, 1883-8014
    • Year and Date
      2017-09-20
    • Related Report
      2017 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] XCSR Learning from Compressed Data Acquired by Deep Neural Network2017

    • Author(s)
      Kazuma Matsumoto, Takato Tatsumi, Hiroyuki Sato, Tim Kovacs, Keiki Takadama
    • Journal Title

      Journal of Advanced Computational Intelligence and Intelligent Informatics

      Volume: 21 Issue: 5 Pages: 856-867

    • DOI

      10.20965/jaciii.2017.p0856

    • NAID

      130007520187

    • ISSN
      1343-0130, 1883-8014
    • Year and Date
      2017-09-20
    • Related Report
      2017 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Approach to Clustering with Variance-Based XCS2017

    • Author(s)
      Zhang, C., Tatsumi, T., Nakata, M., and Takadama, K.
    • Journal Title

      Journal of Advanced Computational Intelligence and Intelligent Informatics

      Volume: 21 Issue: 5 Pages: 885-894

    • DOI

      10.20965/jaciii.2017.p0885

    • NAID

      130007520186

    • ISSN
      1343-0130, 1883-8014
    • Year and Date
      2017-09-20
    • Related Report
      2017 Annual Research Report
    • Peer Reviewed / Open Access
  • [Presentation] 不確実性を伴うデータを分類するルール獲得に向けた正確性によるルール選択メカニズムの設計2019

    • Author(s)
      辰巳 嵩豊, 佐藤 寛之, 髙玉 圭樹
    • Organizer
      進化計算学会進化計算シンポジウム
    • Related Report
      2019 Annual Research Report
  • [Presentation] Comparison of Statistical Table- and Non-Statistical Table-based XCS in Noisy Environments2019

    • Author(s)
      Takato Tatsumi, Keiki Takadama
    • Organizer
      IEEE Congress on Evolutionary Computation
    • Related Report
      2019 Annual Research Report
    • Int'l Joint Research
  • [Presentation] XCS-CR for handling input, output, and reward noise2019

    • Author(s)
      Takato Tatsumi, Keiki Takadama
    • Organizer
      International Workshop on Learning Classifier Systems on Genetic and Evolutionary Computation Conference
    • Related Report
      2019 Annual Research Report
    • Int'l Joint Research
  • [Presentation] XCS-CR: Determining Accuracy of Classifier by Its Collective Reward in Action Set Toward Environment with Action Noise2018

    • Author(s)
      Takato Tatsumi, Tim Kovacs, Keiki Takadama
    • Organizer
      International Workshop on Learning Classifier Systems on Genetic and Evolutionary Computation Conference
    • Related Report
      2018 Annual Research Report
    • Int'l Joint Research
  • [Presentation] XCS for Missing Attributes in Data2018

    • Author(s)
      Takato Tatsumi, Keiki Takadama
    • Organizer
      Joint 10th International Conference on Soft Computing and Intelligent Systems and 19th International Symposium on Advanced Intelligent Systems in conjunction with Intelligent Systems Workshop 2018
    • Related Report
      2018 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 連続報酬環境における学習分類子システムの不適切な一般化に関する一考察2017

    • Author(s)
      辰巳 嵩豊, 高玉圭樹
    • Organizer
      計測自動制御学会システム・情報部門学術講演会 2017
    • Related Report
      2017 Annual Research Report
  • [Presentation] データの曖昧性を許容する学習分類子システム:介護データのマイニング2017

    • Author(s)
      藤野貴章, 辰巳嵩豊, 張財立, 松本和馬, 佐藤寛之, 高玉圭樹
    • Organizer
      計測自動制御学会システム・情報部門学術講演会 2017
    • Related Report
      2017 Annual Research Report
  • [Presentation] Automatic adjustment of selection pressure based on range of reward in learning classifier system2017

    • Author(s)
      Takato Tatsumi,Hiroyuki Sato,Keiki Takadama
    • Organizer
      The Genetic and Evolutionary Computation Conference 2017
    • Related Report
      2017 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Applying variance-based Learning Classifier System without Convergence of Reward Estimation into various Reward distribution2017

    • Author(s)
      Takato Tatsumi, Hiroyuki Sato, Tim Kovacs, Keiki Takadama
    • Organizer
      IEEE Congress on Evolutionary Computation 2017
    • Related Report
      2017 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 限られたデータ数における報酬分散に基づくクラスタリング学習分類子システム:介護プランへの応用2017

    • Author(s)
      張 財立, 辰巳 嵩豊, 高玉 圭樹
    • Organizer
      自動制御連合講演会
    • Related Report
      2017 Annual Research Report

URL: 

Published: 2017-05-25   Modified: 2024-03-26  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi