Hello, everyone:<br>
<br>
I am writing a Bayesian learning library using uBLAS. Since there are a
bounch of matrix inverse operations, for the time being, I try to use
the lu_factorize() defined in lu.hpp for small data sets only
(evetually, I will have to use umfpack considering sparsity). Here is
the simple code I used to try lu_factorize():<br>
<br>
#include &lt;boost/numeric/ublas/io.hpp&gt;<br>
#include &lt;boost/numeric/ublas/matrix.hpp&gt;<br>
#include &lt;boost/numeric/ublas/matrix_proxy.hpp&gt;<br>
#include &lt;boost/numeric/ublas/triangular.hpp&gt;<br>
#include &lt;boost/numeric/ublas/lu.hpp&gt;&nbsp; // the header files are complete?<br>
<br>
using std::cout; <br>
using std::endl;<br>
<br>
int main()<br>
{<br>
&nbsp;&nbsp;&nbsp; namespace ublas = boost::numeric::ublas;<br>
&nbsp;&nbsp;&nbsp; using namespace ublas;<br>
<br>
&nbsp;&nbsp;&nbsp; typedef ublas::vector&lt;double&gt; Vector; <br>
&nbsp;&nbsp;&nbsp; typedef ublas::matrix&lt;double&gt; Matrix;<br>
<br>
&nbsp;&nbsp;&nbsp; Matrix m(3,3); <br>
&nbsp;&nbsp; <br>
&nbsp;&nbsp;&nbsp; for (unsigned i = 0; i &lt; 3; ++ i) <br>
&nbsp;&nbsp;&nbsp; &nbsp; &nbsp; for (unsigned j = 0; j &lt; 3; ++ j)<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; m (i, j) = 3 * i + j + 2;<br>
<br>
&nbsp;&nbsp;&nbsp; Matrix eye3 = ublas::identity_matrix&lt;double&gt;(3, 3);<br>
&nbsp;&nbsp;&nbsp; m = m + 3 * eye3;&nbsp; // now m is not singular<br>
<br>
// test lu factorization<br>
&nbsp;&nbsp;&nbsp; permutation_matrix&lt;std::size_t&gt; pm(3);<br>
&nbsp;&nbsp;&nbsp; <br>
&nbsp;&nbsp;&nbsp; cout &lt;&lt;&quot; m&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = &quot; &lt;&lt; m &lt;&lt; endl;<br>
&nbsp;&nbsp;&nbsp; cout &lt;&lt;&quot; pm&nbsp;&nbsp;&nbsp;&nbsp; = &quot; &lt;&lt; pm &lt;&lt; endl;<br>
&nbsp;&nbsp; <br>
&nbsp;&nbsp; lu_factorize(m);<br>
// lu_factorize(m, pm);<br>
<br>
&nbsp; return 0;<br>
}<br>
<br>
What follows is the error message:<br>
<br>
/usr/include/boost/numeric/ublas/lu.hpp: In function `typename M::size_type<br>
&nbsp;&nbsp; boost::numeric::ublas::lu_factorize(M&amp;) [with M = main()::Matrix]':<br>
try5.cpp:34:&nbsp;&nbsp; instantiated from here<br>
/usr/include/boost/numeric/ublas/lu.hpp:103: error: no matching function for<br>
&nbsp;&nbsp; call to `project(boost::numeric::ublas::matrix_column&lt;main()::Matrix&gt;&amp;,<br>
&nbsp;&nbsp; boost::numeric::ublas::basic_range&lt;size_t, ptrdiff_t&gt;)'<br>
try5.cpp:34:&nbsp;&nbsp; instantiated from here<br>
/usr/include/boost/numeric/ublas/lu.hpp:107: error: no matching function for<br>
&nbsp;&nbsp; call to `project(boost::numeric::ublas::matrix_column&lt;main()::Matrix&gt;&amp;,<br>
&nbsp;&nbsp; boost::numeric::ublas::basic_range&lt;size_t, ptrdiff_t&gt;)'<br>
try5.cpp:34:&nbsp;&nbsp; instantiated from here<br>
/usr/include/boost/numeric/ublas/lu.hpp:107: error: no matching function for<br>
&nbsp;&nbsp; call to `project(boost::numeric::ublas::matrix_row&lt;main()::Matrix&gt;&amp;,<br>
&nbsp;&nbsp; boost::numeric::ublas::basic_range&lt;size_t, ptrdiff_t&gt;)'<br>
<br>
The Boost I use is 1.33 (I downloaded the rpm packages from <a href="http://rpm.pbone.net/">http://rpm.pbone.net/</a>, and installed in my suse 9.1 linux box).<br>
<br>
Can anybody give me some hints on this? ( Probably similar questions have been posted here).<br>
<br>
-Best <br>
<br>
Chen<br>
<br>