<?xml version='1.0' encoding='UTF-8'?>
<OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd">
  <responseDate>2026-04-12T09:21:37Z</responseDate>
  <request identifier="oai:ipsj.ixsq.nii.ac.jp:00234912" metadataPrefix="oai_dc" verb="GetRecord">https://ipsj.ixsq.nii.ac.jp/oai</request>
  <GetRecord>
    <record>
      <header>
        <identifier>oai:ipsj.ixsq.nii.ac.jp:00234912</identifier>
        <datestamp>2025-01-19T09:40:23Z</datestamp>
        <setSpec>1164:2735:11468:11669</setSpec>
      </header>
      <metadata>
        <oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns="http://www.w3.org/2001/XMLSchema" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
          <dc:title>学習係数を用いたモデル選択sBICの解析およびその応用</dc:title>
          <dc:title>Analysis of the model selection method sBIC with learning coeﬃcients and its applications</dc:title>
          <dc:creator>山崎, 敬太</dc:creator>
          <dc:creator>大場, 智康</dc:creator>
          <dc:creator>小林, 晴</dc:creator>
          <dc:creator>清水, 恭介</dc:creator>
          <dc:creator>梶, 大介</dc:creator>
          <dc:creator>青柳, 美輝</dc:creator>
          <dc:creator>Keita, Yamazaki</dc:creator>
          <dc:creator>Tomoyasu, Ohba</dc:creator>
          <dc:creator>Haru, Kobayashi</dc:creator>
          <dc:creator>Kyousuke, Shimizu</dc:creator>
          <dc:creator>Daisuke, Kaji</dc:creator>
          <dc:creator>Miki, Aoyagi</dc:creator>
          <dc:subject>情報論的学習理論と機械学習3</dc:subject>
          <dc:description>近年，特異モデルのベイズ学習について，汎化誤差や経験誤差，自由エネルギーなどの漸近挙動を，学習係数やその位数，特異揺らぎを用いて解析できることが証明された [1]．理論値は，学習曲線の特徴を表しており，解析や応用に重要な指標である．特に，学習係数は，代数幾何の分野において，log canonical threshold として定義され，汎化誤差の主要項を表す．本研究では，学習係数を用いたモデル選択の手法である特異ベイズ情報量基準 (sBIC) について考察し，その応用として新型コロナウイルス感染症の重症者数推定のモデルを線形ニューラルネットワークとして知られる縮小ランクモデルを使って構築する．</dc:description>
          <dc:description>In recent years, it has been demonstrated that the asymptotic behaviors of generalization error, empirical error, and free energy in singular model Bayesian learning can be analyzed using learning coeﬃcients, their orders, and singular ﬂuctuations [1]. The theoretical values represent the characteristics of learning curves and serve as important indicators for analysis and application. Particularly, the learning coeﬃcient is deﬁned as the log canonical threshold in the ﬁeld of algebraic geometry, serving as an indicator of generalization error. In this study, we examine the Singular Bayesian Information Criterion (sBIC), a method for model selection using learning coeﬃcients. As an application, we construct a model for estimating the number of severe cases of COVID-19 using a reduced-rank model.</dc:description>
          <dc:description>technical report</dc:description>
          <dc:publisher>情報処理学会</dc:publisher>
          <dc:date>2024-06-13</dc:date>
          <dc:format>application/pdf</dc:format>
          <dc:identifier>研究報告数理モデル化と問題解決（MPS）</dc:identifier>
          <dc:identifier>25</dc:identifier>
          <dc:identifier>2024-MPS-148</dc:identifier>
          <dc:identifier>1</dc:identifier>
          <dc:identifier>7</dc:identifier>
          <dc:identifier>2188-8833</dc:identifier>
          <dc:identifier>AN10505667</dc:identifier>
          <dc:identifier>https://ipsj.ixsq.nii.ac.jp/record/234912/files/IPSJ-MPS24148025.pdf</dc:identifier>
          <dc:language>jpn</dc:language>
        </oai_dc:dc>
      </metadata>
    </record>
  </GetRecord>
</OAI-PMH>
