<rss xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" version="2.0">
  <channel>
    <title>MaplePrimes - answers and comments on Question, Error, (in Statistics:-Fit) SVD of estimated Jacobian could not be computed</title>
    <link>http://www.mapleprimes.com/questions/139234-Error-in-StatisticsFit-SVD-Of-Estimated</link>
    <language>en-us</language>
    <copyright>2026 Maplesoft, A Division of Waterloo Maple Inc.</copyright>
    <generator>Maplesoft Document System</generator>
    <lastBuildDate>Tue, 09 Jun 2026 18:20:46 GMT</lastBuildDate>
    <pubDate>Tue, 09 Jun 2026 18:20:46 GMT</pubDate>
    <itunes:subtitle />
    <itunes:summary />
    <description>The latest answers and comments added to the Question, Error, (in Statistics:-Fit) SVD of estimated Jacobian could not be computed</description>
    <image>
      <url>http://www.mapleprimes.com/images/mapleprimeswhite.jpg</url>
      <title>MaplePrimes - answers and comments on Question, Error, (in Statistics:-Fit) SVD of estimated Jacobian could not be computed</title>
      <link>http://www.mapleprimes.com/questions/139234-Error-in-StatisticsFit-SVD-Of-Estimated</link>
    </image>
    <item>
      <title>Provide initial values</title>
      <link>http://www.mapleprimes.com/questions/139234-Error-in-StatisticsFit-SVD-Of-Estimated?ref=Feed:MaplePrimes:Error, (in Statistics:-Fit) SVD of estimated Jacobian could not be computed:Comments#answer139254</link>
      <itunes:summary>&lt;p&gt;Hi GaoCG,&lt;/p&gt;
&lt;p&gt;You'll need to provide initial values for your parameters so that Maple has a starting point for iterating towards a locally optimal solution. You can do this with the initialvalues option explained on the&amp;nbsp;&lt;a href="http://www.maplesoft.com/support/help/view.aspx?path=Statistics/Fit"&gt;Statistics/Fit help page&lt;/a&gt;. Also, there seems to be an extra parenthesis in the call to Fit; I assume that's the one just before ', X, Y, sigma', right?&lt;/p&gt;
&lt;p&gt;It's not so easy to find values of the parameters that lead to numbers that are anywhere near reasonable. Maybe you know the actual value of the parameters for some other instance of the model? Since there's a factor of sigma^2-(650*10^6)^2 in the denominator, where sigma^2 is on the order of a few hundred, this is a big negative number; since we need to find values that are positive, we need to compensate for that. One option would be to set A to a big negative number.&lt;/p&gt;
&lt;p&gt;Another thing to remember is that there might be bad cancellation going on: because you're subtracting such a big number from sigma^2, if you want to see the effect of sigma on the result, you'll need sufficiently many digits to represent it accurately. It might be useful to increase Digits to a higher number, like 25.&lt;/p&gt;
&lt;p&gt;Hope this helps,&lt;/p&gt;
&lt;p&gt;Erik Postma&lt;br&gt;Maplesoft.&lt;/p&gt;</itunes:summary>
      <description>&lt;p&gt;Hi GaoCG,&lt;/p&gt;
&lt;p&gt;You'll need to provide initial values for your parameters so that Maple has a starting point for iterating towards a locally optimal solution. You can do this with the initialvalues option explained on the&amp;nbsp;&lt;a href="http://www.maplesoft.com/support/help/view.aspx?path=Statistics/Fit"&gt;Statistics/Fit help page&lt;/a&gt;. Also, there seems to be an extra parenthesis in the call to Fit; I assume that's the one just before ', X, Y, sigma', right?&lt;/p&gt;
&lt;p&gt;It's not so easy to find values of the parameters that lead to numbers that are anywhere near reasonable. Maybe you know the actual value of the parameters for some other instance of the model? Since there's a factor of sigma^2-(650*10^6)^2 in the denominator, where sigma^2 is on the order of a few hundred, this is a big negative number; since we need to find values that are positive, we need to compensate for that. One option would be to set A to a big negative number.&lt;/p&gt;
&lt;p&gt;Another thing to remember is that there might be bad cancellation going on: because you're subtracting such a big number from sigma^2, if you want to see the effect of sigma on the result, you'll need sufficiently many digits to represent it accurately. It might be useful to increase Digits to a higher number, like 25.&lt;/p&gt;
&lt;p&gt;Hope this helps,&lt;/p&gt;
&lt;p&gt;Erik Postma&lt;br&gt;Maplesoft.&lt;/p&gt;</description>
      <guid>139254</guid>
      <pubDate>Wed, 07 Nov 2012 00:38:10 Z</pubDate>
      <itunes:author>epostma</itunes:author>
      <author>epostma</author>
    </item>
  </channel>
</rss>