<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Math on Klaus K. Holst</title>
    <link>https://holst.it/tags/math/</link>
    <description>Recent content in Math on Klaus K. Holst</description>
    <generator>Hugo -- 0.148.2</generator>
    <language>en</language>
    <lastBuildDate>Fri, 19 Jun 2020 20:20:00 +0200</lastBuildDate>
    <atom:link href="https://holst.it/tags/math/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Nonlinear latent variable models</title>
      <link>https://holst.it/posts/nsem/</link>
      <pubDate>Fri, 19 Jun 2020 20:20:00 +0200</pubDate>
      <guid>https://holst.it/posts/nsem/</guid>
      <description>&lt;p&gt;ML-inference in non-linear SEMs is complex. Computational intensive
methods based on numerical integration are needed and results are
sensitive to distributional assumptions.&lt;/p&gt;
&lt;p&gt;In a recent paper:
&lt;strong&gt;A two-stage estimation procedure for non-linear structural equation models&lt;/strong&gt;
by &lt;em&gt;Klaus Kähler Holst &amp;amp; Esben Budtz-Jørgensen (&lt;a href=&#34;https://doi.org/10.1093/biostatistics/kxy082&#34; target=&#34;_blank&#34;&gt;https://doi.org/10.1093/biostatistics/kxy082&lt;/a&gt;)&lt;/em&gt;,
we consider two-stage estimators as a &lt;strong&gt;computationally simple alternative to MLE&lt;/strong&gt;.
Here both steps are based on linear models: first we predict the non-linear
terms and then these are related to latent outcomes in the second step.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
