<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>機械学習 on matsueushi</title>
    <link>https://matsueushi.github.io/tags/%E6%A9%9F%E6%A2%B0%E5%AD%A6%E7%BF%92/</link>
    <description>Recent content in 機械学習 on matsueushi</description>
    <generator>Hugo</generator>
    <language>ja</language>
    <copyright>Copyright © 2019–2026 matsueushi</copyright>
    <lastBuildDate>Thu, 02 Feb 2023 22:52:35 +0900</lastBuildDate>
    <atom:link href="https://matsueushi.github.io/tags/%E6%A9%9F%E6%A2%B0%E5%AD%A6%E7%BF%92/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>MlflowでCatBoostのメトリクスをロギングする</title>
      <link>https://matsueushi.github.io/ds/catboost-mlflow/</link>
      <pubDate>Thu, 02 Feb 2023 22:52:35 +0900</pubDate>
      <guid>https://matsueushi.github.io/ds/catboost-mlflow/</guid>
      <description>&lt;p&gt;LightGBM, XGBoostであれば、途中の学習メトリクスをMlflowで記録するとき、&lt;code&gt;mlflow.lightgbm.autolog()&lt;/code&gt; や &lt;code&gt;mlflow.xgboost.autolog()&lt;/code&gt; を呼び出すだけで記録してくれるので非常に楽である。一方、CatBoost にはオートログの機能はない。&lt;/p&gt;</description>
    </item>
    <item>
      <title>Flux.jl v0.10.0でDCGANを動かす(CUDA環境)</title>
      <link>https://matsueushi.github.io/ds/fluxjl-dcgan/</link>
      <pubDate>Sun, 01 Dec 2019 18:07:13 -0500</pubDate>
      <guid>https://matsueushi.github.io/ds/fluxjl-dcgan/</guid>
      <description>&lt;p&gt;2020/3/8 追記: model-zoo(&lt;a href=&#34;https://github.com/FluxML/model-zoo&#34;&gt;https://github.com/FluxML/model-zoo&lt;/a&gt;) にDCGANのモデルが入りました。&lt;/p&gt;&#xA;&lt;hr&gt;&#xA;&lt;p&gt;2019/11/29にJuliaの機械学習ライブラリ&lt;a href=&#34;https://fluxml.ai/Flux.jl/stable/&#34;&gt;Flux.jl&lt;/a&gt;の&lt;a href=&#34;https://github.com/FluxML/Flux.jl/releases/tag/v0.10.0&#34;&gt;v0.10.0&lt;/a&gt;がリリースされた。&#xA;もともとv0.9.0でDCGANのMNISTデータセットから手書き文字画像生成モデルを作成して、今回の変更に合わせてv0.10.0で動かしたのだが、&#xA;ここに至るまで色々と苦戦したので、v0.10.0の主な変更点や、自分がつまづいた点を書いておく。&lt;/p&gt;</description>
    </item>
    <item>
      <title>Juliaのガウス過程ライブラリGaussianProcesses.jlを使う(ハイパーパラメーター推定)</title>
      <link>https://matsueushi.github.io/ds/gaussianprocess-jl-2/</link>
      <pubDate>Sun, 04 Aug 2019 11:04:13 -0400</pubDate>
      <guid>https://matsueushi.github.io/ds/gaussianprocess-jl-2/</guid>
      <description>&lt;p&gt;&lt;a href=&#34;../gaussianprocess-jl/&#34;&gt;前回&lt;/a&gt; の続き。GaussianProcesses.jlを使ってガウス過程回帰のハイパーパラメータ推定を行ってみる。&#xA;「&lt;a href=&#34;http://chasen.org/~daiti-m/gpbook/&#34;&gt;ガウス過程と機械学習&lt;/a&gt;」の図3.20(a)を例とする。&lt;/p&gt;&#xA;&lt;div class=&#34;highlight&#34;&gt;&lt;pre tabindex=&#34;0&#34; style=&#34;background-color:#f7f7f7;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;&#34;&gt;&lt;code class=&#34;language-julia&#34; data-lang=&#34;julia&#34;&gt;&lt;span style=&#34;display:flex;&#34;&gt;&lt;span&gt;xxs &lt;span style=&#34;color:#0550ae&#34;&gt;=&lt;/span&gt; &lt;span style=&#34;color:#1f2328&#34;&gt;[&lt;/span&gt;&lt;span style=&#34;color:#0550ae&#34;&gt;-&lt;/span&gt;&lt;span style=&#34;color:#0550ae&#34;&gt;0.5&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;0.5&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;1&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;1.4&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;3&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;]&lt;/span&gt;&#xA;&lt;/span&gt;&lt;/span&gt;&lt;span style=&#34;display:flex;&#34;&gt;&lt;span&gt;yys &lt;span style=&#34;color:#0550ae&#34;&gt;=&lt;/span&gt; &lt;span style=&#34;color:#1f2328&#34;&gt;[&lt;/span&gt;&lt;span style=&#34;color:#0550ae&#34;&gt;0.7&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;1.8&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;1.7&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;2.3&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; &lt;span style=&#34;color:#0550ae&#34;&gt;1&lt;/span&gt;&lt;span style=&#34;color:#1f2328&#34;&gt;]&lt;/span&gt;&#xA;&lt;/span&gt;&lt;/span&gt;&lt;span style=&#34;display:flex;&#34;&gt;&lt;span&gt;&#xA;&lt;/span&gt;&lt;/span&gt;&lt;span style=&#34;display:flex;&#34;&gt;&lt;span&gt;scatter&lt;span style=&#34;color:#1f2328&#34;&gt;(&lt;/span&gt;xxs&lt;span style=&#34;color:#1f2328&#34;&gt;,&lt;/span&gt; yys&lt;span style=&#34;color:#1f2328&#34;&gt;)&lt;/span&gt;&#xA;&lt;/span&gt;&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;&lt;/div&gt;&lt;svg xmlns=&#34;http://www.w3.org/2000/svg&#34; xmlns:xlink=&#34;http://www.w3.org/1999/xlink&#34; width=&#34;600&#34; height=&#34;400&#34; viewBox=&#34;0 0 2400 1600&#34;&gt;&#xA;&lt;defs&gt;&#xA;  &lt;clipPath id=&#34;clip8400&#34;&gt;&#xA;    &lt;rect x=&#34;0&#34; y=&#34;0&#34; width=&#34;2400&#34; height=&#34;1600&#34;/&gt;&#xA;  &lt;/clipPath&gt;&#xA;&lt;/defs&gt;&#xA;&lt;polygon clip-path=&#34;url(#clip8400)&#34; points=&#34;&#xA;0,1600 2400,1600 2400,0 0,0 &#xA;  &#34; fill=&#34;#ffffff&#34; fill-rule=&#34;evenodd&#34; fill-opacity=&#34;1&#34;/&gt;&#xA;&lt;defs&gt;&#xA;  &lt;clipPath id=&#34;clip8401&#34;&gt;&#xA;    &lt;rect x=&#34;480&#34; y=&#34;0&#34; width=&#34;1681&#34; height=&#34;1600&#34;/&gt;&#xA;  &lt;/clipPath&gt;&#xA;&lt;/defs&gt;&#xA;&lt;polygon clip-path=&#34;url(#clip8400)&#34; points=&#34;&#xA;161.394,1503.47 2321.26,1503.47 2321.26,47.2441 161.394,47.2441 &#xA;  &#34; fill=&#34;#ffffff&#34; fill-rule=&#34;evenodd&#34; fill-opacity=&#34;1&#34;/&gt;&#xA;&lt;defs&gt;&#xA;  &lt;clipPath id=&#34;clip8402&#34;&gt;&#xA;    &lt;rect x=&#34;161&#34; y=&#34;47&#34; width=&#34;2161&#34; height=&#34;1457&#34;/&gt;&#xA;  &lt;/clipPath&gt;&#xA;&lt;/defs&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  513.609,1503.47 513.609,47.2441 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  1095.78,1503.47 1095.78,47.2441 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  1677.96,1503.47 1677.96,47.2441 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  2260.13,1503.47 2260.13,47.2441 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  161.394,1290.54 2321.26,1290.54 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  161.394,1032.95 2321.26,1032.95 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  161.394,775.359 2321.26,775.359 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  161.394,517.771 2321.26,517.771 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8402)&#34; style=&#34;stroke:#000000; stroke-width:2; stroke-opacity:0.1; fill:none&#34; points=&#34;&#xA;  161.394,260.183 2321.26,260.183 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,1503.47 2321.26,1503.47 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,1503.47 161.394,47.2441 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  513.609,1503.47 513.609,1481.63 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  1095.78,1503.47 1095.78,1481.63 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  1677.96,1503.47 1677.96,1481.63 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  2260.13,1503.47 2260.13,1481.63 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,1290.54 193.792,1290.54 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,1032.95 193.792,1032.95 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,775.359 193.792,775.359 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,517.771 193.792,517.771 &#xA;  &#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  161.394,260.183 193.792,260.183 &#xA;  &#34;/&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:middle;&#34; transform=&#34;rotate(0, 513.609, 1557.47)&#34; x=&#34;513.609&#34; y=&#34;1557.47&#34;&gt;0&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:middle;&#34; transform=&#34;rotate(0, 1095.78, 1557.47)&#34; x=&#34;1095.78&#34; y=&#34;1557.47&#34;&gt;1&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:middle;&#34; transform=&#34;rotate(0, 1677.96, 1557.47)&#34; x=&#34;1677.96&#34; y=&#34;1557.47&#34;&gt;2&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:middle;&#34; transform=&#34;rotate(0, 2260.13, 1557.47)&#34; x=&#34;2260.13&#34; y=&#34;1557.47&#34;&gt;3&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:end;&#34; transform=&#34;rotate(0, 137.394, 1308.04)&#34; x=&#34;137.394&#34; y=&#34;1308.04&#34;&gt;0.9&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:end;&#34; transform=&#34;rotate(0, 137.394, 1050.45)&#34; x=&#34;137.394&#34; y=&#34;1050.45&#34;&gt;1.2&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:end;&#34; transform=&#34;rotate(0, 137.394, 792.859)&#34; x=&#34;137.394&#34; y=&#34;792.859&#34;&gt;1.5&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:end;&#34; transform=&#34;rotate(0, 137.394, 535.271)&#34; x=&#34;137.394&#34; y=&#34;535.271&#34;&gt;1.8&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:end;&#34; transform=&#34;rotate(0, 137.394, 277.683)&#34; x=&#34;137.394&#34; y=&#34;277.683&#34;&gt;2.1&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#000000; stroke:none; fill-opacity:1&#34; cx=&#34;222.522&#34; cy=&#34;1462.26&#34; r=&#34;18&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#009af9; stroke:none; fill-opacity:1&#34; cx=&#34;222.522&#34; cy=&#34;1462.26&#34; r=&#34;14&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#000000; stroke:none; fill-opacity:1&#34; cx=&#34;804.696&#34; cy=&#34;517.771&#34; r=&#34;18&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#009af9; stroke:none; fill-opacity:1&#34; cx=&#34;804.696&#34; cy=&#34;517.771&#34; r=&#34;14&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#000000; stroke:none; fill-opacity:1&#34; cx=&#34;1095.78&#34; cy=&#34;603.634&#34; r=&#34;18&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#009af9; stroke:none; fill-opacity:1&#34; cx=&#34;1095.78&#34; cy=&#34;603.634&#34; r=&#34;14&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#000000; stroke:none; fill-opacity:1&#34; cx=&#34;1328.65&#34; cy=&#34;88.4582&#34; r=&#34;18&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#009af9; stroke:none; fill-opacity:1&#34; cx=&#34;1328.65&#34; cy=&#34;88.4582&#34; r=&#34;14&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#000000; stroke:none; fill-opacity:1&#34; cx=&#34;2260.13&#34; cy=&#34;1204.67&#34; r=&#34;18&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8402)&#34; style=&#34;fill:#009af9; stroke:none; fill-opacity:1&#34; cx=&#34;2260.13&#34; cy=&#34;1204.67&#34; r=&#34;14&#34;/&gt;&#xA;&lt;polygon clip-path=&#34;url(#clip8400)&#34; points=&#34;&#xA;1958.43,251.724 2249.26,251.724 2249.26,130.764 1958.43,130.764 &#xA;  &#34; fill=&#34;#ffffff&#34; fill-rule=&#34;evenodd&#34; fill-opacity=&#34;1&#34;/&gt;&#xA;&lt;polyline clip-path=&#34;url(#clip8400)&#34; style=&#34;stroke:#000000; stroke-width:4; stroke-opacity:1; fill:none&#34; points=&#34;&#xA;  1958.43,251.724 2249.26,251.724 2249.26,130.764 1958.43,130.764 1958.43,251.724 &#xA;  &#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8400)&#34; style=&#34;fill:#000000; stroke:none; fill-opacity:1&#34; cx=&#34;2066.43&#34; cy=&#34;191.244&#34; r=&#34;25&#34;/&gt;&#xA;&lt;circle clip-path=&#34;url(#clip8400)&#34; style=&#34;fill:#009af9; stroke:none; fill-opacity:1&#34; cx=&#34;2066.43&#34; cy=&#34;191.244&#34; r=&#34;21&#34;/&gt;&#xA;&lt;g clip-path=&#34;url(#clip8400)&#34;&gt;&#xA;&lt;text style=&#34;fill:#000000; fill-opacity:1; font-family:Arial,Helvetica Neue,Helvetica,sans-serif; font-size:48px; text-anchor:start;&#34; transform=&#34;rotate(0, 2150.43, 208.744)&#34; x=&#34;2150.43&#34; y=&#34;208.744&#34;&gt;y1&lt;/text&gt;&#xA;&lt;/g&gt;&#xA;&lt;/svg&gt;&#xA;&lt;p&gt;平均0のガウスカーネルのガウス過程に(ノイズ項込みで)フィッティングさせるのは非常に簡単で、&lt;code&gt;optimize!&lt;/code&gt; を呼べば良いだけ。&lt;/p&gt;</description>
    </item>
    <item>
      <title>Juliaのガウス過程ライブラリGaussianProcesses.jlを使う(基本編)</title>
      <link>https://matsueushi.github.io/ds/gaussianprocess-jl/</link>
      <pubDate>Wed, 17 Jul 2019 21:22:00 -0400</pubDate>
      <guid>https://matsueushi.github.io/ds/gaussianprocess-jl/</guid>
      <description>&lt;p&gt;Juliaのガウス過程ライブラリ &lt;a href=&#34;https://github.com/STOR-i/GaussianProcesses.jl&#34;&gt;GaussianProcesses.jl&lt;/a&gt; を触ってみる。&#xA;チュートリアルはまだ未整備のようなので、&lt;a href=&#34;https://github.com/STOR-i/GaussianProcesses.jl/tree/master/notebooks&#34;&gt;サンプルノートブック&lt;/a&gt;と&lt;a href=&#34;http://chasen.org/~daiti-m/gpbook/&#34;&gt;ガウス過程と機械学習&lt;/a&gt;」などを参考にしながらやっていく。&#xA;Juliaのはv1.1.0, GaussianProcesses.jlはv0.9.0を使っている。&lt;/p&gt;</description>
    </item>
    <item>
      <title>Juliaでガウス過程を実装&amp;パラメーター推定</title>
      <link>https://matsueushi.github.io/ds/gp-parameter-estimation/</link>
      <pubDate>Sat, 08 Jun 2019 20:08:12 -0400</pubDate>
      <guid>https://matsueushi.github.io/ds/gp-parameter-estimation/</guid>
      <description>&lt;p&gt;「ガウス過程と機械学習」を3章まで読み終えたので、復習を兼ねてJulia(1.1.0)でガウス過程を実装し、&#xA;カーネルのハイパーパラメーターをOptim.jlで推定するところまでをまとめる。数学的に細かい内容は本を読んで欲しい。&#xA;図3.23の陸上男子100mの世界記録の回帰モデルを作成することを今回の目標とする。&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
