testing output

This commit is contained in:
Fabrizio Riguzzi 2013-09-17 14:10:08 +02:00
parent b9d4868e37
commit 3659f9036c
4 changed files with 54 additions and 27 deletions

View File

@ -7,7 +7,7 @@
<meta name="originator" content="TeX4ht (http://www.cse.ohio-state.edu/~gurari/TeX4ht/)">
<!-- html -->
<meta name="src" content="manual.tex">
<meta name="date" content="2013-09-17 13:02:00">
<meta name="date" content="2013-09-17 14:09:00">
<link rel="stylesheet" type="text/css" href="manual.css">
</head><body
>
@ -1349,7 +1349,7 @@ class="cmtt-10">ll</span></span></span>, valid for SLIPCOVER):
determines the score function for refinement: if set to <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">ll</span></span></span>, log likelihood is
used, if set to <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">aucpr</span></span></span>, the area under the precision-recall curve is used.</li></ul>
class="cmtt-10">aucpr</span></span></span>, the area under the Precision-Recall curve is used.</li></ul>
<!--l. 635--><p class="noindent" >
<h4 class="subsectionHead"><span class="titlemark">5.3 </span> <a
@ -1410,21 +1410,22 @@ class="cmtt-10">slipcover.pl </span>with
?:-&#x00A0;sl(stem).
</div>
<!--l. 675--><p class="nopar" >
<!--l. 678--><p class="noindent" >
<h4 class="subsectionHead"><span class="titlemark">5.4 </span> <a
id="x1-140005.4"></a>Testing</h4>
<!--l. 677--><p class="noindent" >To test the theories learned, load <span
<!--l. 679--><p class="noindent" >To test the theories learned, load <span
class="cmtt-10">test.pl </span>with
<div class="verbatim" id="verbatim-64">
?:-&#x00A0;use_module(library(&#8217;cplint/test&#8217;)).
</div>
<!--l. 680--><p class="nopar" > and call
<!--l. 682--><p class="nopar" > and call
<div class="verbatim" id="verbatim-65">
?:-&#x00A0;main([&#x003C;stem_fold1&#x003E;,...,&#x003C;stem_foldn&#x003E;],[&#x003C;testing_set_fold1&#x003E;,...,
&#x00A0;<br />&#x00A0;&#x00A0;&#x003C;testing_set_foldn&#x003E;]).
</div>
<!--l. 685--><p class="nopar" > For example, if you want to test the theory in <span class="obeylines-h"><span class="verb"><span
<!--l. 687--><p class="nopar" > For example, if you want to test the theory in <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">ai_train.rules</span></span></span> on the set <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">ai.kb</span></span></span>,
you can call
@ -1432,7 +1433,7 @@ you can call
<div class="verbatim" id="verbatim-66">
?:-&#x00A0;main([ai_train],[ai]).
</div>
<!--l. 689--><p class="nopar" > The testing program has the following parameter:
<!--l. 691--><p class="nopar" > The testing program has the following parameter:
<ul class="itemize1">
<li class="itemize"><span class="obeylines-h"><span class="verb"><span
class="cmtt-10">neg_ex</span></span></span> (values: <span class="obeylines-h"><span class="verb"><span
@ -1450,10 +1451,26 @@ class="cmtt-10">cw</span></span></span>, the negative examples are generated
predicates that are not positive examples. The set of all atoms is obtained
by collecting the set of constants for each type of the arguments of the
target predicate.</li></ul>
<!--l. 696--><p class="noindent" >
<!--l. 697--><p class="noindent" >The testing program produces the following output in the current folder:
<ul class="itemize1">
<li class="itemize"><span class="obeylines-h"><span class="verb"><span
class="cmtt-10">cll.pl</span></span></span>: for each fold, the list of examples orderd by their probability of
being true
</li>
<li class="itemize"><span class="obeylines-h"><span class="verb"><span
class="cmtt-10">areas.csv</span></span></span>: the areas under the Precision-Recall curve and the Receiver
Operating Characteristic curve
</li>
<li class="itemize"><span class="obeylines-h"><span class="verb"><span
class="cmtt-10">curve_roc.m</span></span></span>: a Matlab file for plotting the Receiver Operating
Characteristic curve
</li>
<li class="itemize"><span class="obeylines-h"><span class="verb"><span
class="cmtt-10">curve_pr.m</span></span></span>: a Matlab file for plotting the Precision-Recall curve</li></ul>
<!--l. 706--><p class="noindent" >
<h4 class="subsectionHead"><span class="titlemark">5.5 </span> <a
id="x1-150005.5"></a>Learning Examples</h4>
<!--l. 697--><p class="noindent" >The subfolders <span class="obeylines-h"><span class="verb"><span
<!--l. 707--><p class="noindent" >The subfolders <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">em</span></span></span>, <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">rib</span></span></span>, <span class="obeylines-h"><span class="verb"><span
class="cmtt-10">slipcase</span></span></span> and <span class="obeylines-h"><span class="verb"><span
@ -1461,27 +1478,27 @@ class="cmtt-10">slipcover</span></span></span> of the <span class="obeylines-h">
class="cmtt-10">packages/cplint</span></span></span> folder in
Yap git distribution contain examples of input and output files for the learning
algorithms.
<!--l. 700--><p class="noindent" >
<!--l. 710--><p class="noindent" >
<h3 class="sectionHead"><span class="titlemark">6 </span> <a
id="x1-160006"></a>License</h3>
<!--l. 705--><p class="noindent" ><span
<!--l. 715--><p class="noindent" ><span
class="cmtt-10">cplint</span>, as Yap, follows the Artistic License 2.0 that you can find in Yap CVS root
dir. The copyright is by Fabrizio Riguzzi.
<!--l. 708--><p class="indent" > The modules in the approx subdirectory use SimplecuddLPADs, a modification of
<!--l. 718--><p class="indent" > The modules in the approx subdirectory use SimplecuddLPADs, a modification of
the <a
href="http://dtai.cs.kuleuven.be/problog/download.html" > Simplecudd </a> library whose copyright is by Katholieke Universiteit Leuven and
that follows the Artistic License 2.0.
<!--l. 711--><p class="indent" > Some modules use the library <a
<!--l. 721--><p class="indent" > Some modules use the library <a
href="http://vlsi.colorado.edu/~fabio/" > CUDD </a> for manipulating BDDs that is included in
glu. For the use of CUDD, the following license must be accepted:
<!--l. 716--><p class="indent" > Copyright (c) 1995-2004, Regents of the University of Colorado
<!--l. 718--><p class="indent" > All rights reserved.
<!--l. 720--><p class="indent" > Redistribution and use in source and binary forms, with or without modification,
<!--l. 726--><p class="indent" > Copyright (c) 1995-2004, Regents of the University of Colorado
<!--l. 728--><p class="indent" > All rights reserved.
<!--l. 730--><p class="indent" > Redistribution and use in source and binary forms, with or without modification,
are permitted provided that the following conditions are met:
<ul class="itemize1">
<li class="itemize">Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.
</li>
<li class="itemize">Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
@ -1490,7 +1507,7 @@ are permitted provided that the following conditions are met:
<li class="itemize">Neither the name of the University of Colorado nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.</li></ul>
<!--l. 737--><p class="noindent" >THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS <br
<!--l. 747--><p class="noindent" >THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS <br
class="newline" />AND CONTRIBUTORS &#8221;AS IS&#8221; AND ANY EXPRESS OR IMPLIED
WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
@ -1504,7 +1521,7 @@ class="newline" />AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
<!--l. 751--><p class="indent" > <span
<!--l. 761--><p class="indent" > <span
class="cmtt-10">lpad.pl</span>, <span
class="cmtt-10">semlpad.pl </span>and <span
class="cmtt-10">cpl.pl </span>are based on the SLG system by Weidong
@ -1512,6 +1529,7 @@ Chen and <a
href="http://www.cs.sunysb.edu/~warren/" > David Scott Warren </a>, Copyright (C) 1993 Southern Methodist University,
1993 SUNY at Stony Brook, see the file COYPRIGHT_SLG for detailed information
on this copyright.
<!--l. 1--><p class="noindent" >
<h3 class="likesectionHead"><a
id="x1-170006"></a>References</h3>
@ -1532,7 +1550,6 @@ class="cmti-10">Proceedings of the 26th Italian</span>
class="cmti-10">Conference on Computational Logic (CILC2011), Pescara, Italy, 31 August</span>
<span
class="cmti-10">31-2 September, 2011</span>, 2011.
</p>
<p class="bibitem" ><span class="biblabel">
[3]<span class="bibsp">&#x00A0;&#x00A0;&#x00A0;</span></span><a
@ -1571,6 +1588,7 @@ class="cmti-10">Programming</span>, 2013.
class="cmti-10">Inductive Logic Programming (ILP 2004), Work in Progress</span>
<span
class="cmti-10">Track</span>, 2004.
</p>
<p class="bibitem" ><span class="biblabel">
[8]<span class="bibsp">&#x00A0;&#x00A0;&#x00A0;</span></span><a
@ -1589,7 +1607,6 @@ class="cmti-10">LNCS</span>, pages 30&#8211;37. Springer, 2011.
delaying for general logic programs. <span
class="cmti-10">Journal of the ACM</span>, 43(1):20&#8211;74, 1996.
</p>
<p class="bibitem" ><span class="biblabel">
[10]<span class="bibsp">&#x00A0;&#x00A0;&#x00A0;</span></span><a
id="XDBLP:journals/ai/Cohen95"></a>William&#x00A0;W. Cohen. Pac-learning non-recursive prolog clauses. <span
@ -1632,6 +1649,7 @@ class="cmti-10">Research</span>, 3:679&#8211;707, December 2002.
id="XDBLP:journals/ai/Poole97"></a>David Poole. The independent choice logic for modelling multiple agents
under uncertainty. <span
class="cmti-10">Artificial Intelligence</span>, 94(1-2):7&#8211;56, 1997.
</p>
<p class="bibitem" ><span class="biblabel">
[16]<span class="bibsp">&#x00A0;&#x00A0;&#x00A0;</span></span><a
@ -1648,7 +1666,6 @@ class="cmti-10">LNAI</span>, pages 109&#8211;120. Springer, 2007.
class="cmti-10">Proceedings of the 14th RCRA workshop Experimental Evaluation of</span>
<span
class="cmti-10">Algorithms for Solving Problems with Combinatorial Explosion</span>, 2007.
</p>
<p class="bibitem" ><span class="biblabel">
[18]<span class="bibsp">&#x00A0;&#x00A0;&#x00A0;</span></span><a
@ -1693,6 +1710,7 @@ class="cmti-10">Proceedings of the 10th</span>
class="cmti-10">European Conference on Logics in Artificial Intelligence</span>, LNAI. Springer,
September 2006.
</p>
<p class="bibitem" ><span class="biblabel">
[23]<span class="bibsp">&#x00A0;&#x00A0;&#x00A0;</span></span><a
id="XDBLP:journals/tplp/VennekensDB09"></a>J.&#x00A0;Vennekens, Marc Denecker, and Maurice Bruynooghe. CP-logic:
@ -1713,7 +1731,6 @@ class="cmti-10">International Conference on Logic</span>
<span
class="cmti-10">Programming</span>, volume 3131 of <span
class="cmti-10">LNCS</span>, pages 195&#8211;209. Springer, 2004.
</p>
</div>

Binary file not shown.

View File

@ -630,7 +630,7 @@ maximum number of theory search iterations
\item \verb|maxdepth_var| (values: integer, default value: 2, valid for SLIPCOVER): maximum depth of
variables in clauses (as defined in \cite{DBLP:journals/ai/Cohen95}).
\item \verb|score| (values: \verb|ll|, \verb|aucpr|, default value \verb|ll|, valid for SLIPCOVER): determines the score function for refinement: if set to \verb|ll|, log likelihood is used, if set to \verb|aucpr|, the area under the
precision-recall curve is used.
Precision-Recall curve is used.
\end{itemize}
\subsection{Commands}
To execute CEM, load \texttt{em.pl} with
@ -673,6 +673,8 @@ and call
\begin{verbatim}
?:- sl(stem).
\end{verbatim}
\subsection{Testing}
To test the theories learned, load \texttt{test.pl} with
\begin{verbatim}
@ -692,6 +694,14 @@ The testing program has the following parameter:
\item \verb|neg_ex| (values: \verb|given|, \verb|cw|, default value: \verb|cw|): if set to \verb|given|, the negative examples
are taken from \verb|<testing_set_foldi>.kb|, i.e., those example \verb|ex| stored as \verb|neg(ex)|; if set to \verb|cw|, the negative examples are generated according to the closed world assumption, i.e., all atoms for target predicates that are not positive examples. The set of all atoms is obtained by collecting the set of constants for each type of the arguments of the target predicate.
\end{itemize}
The testing program produces the following output in the current folder:
\begin{itemize}
\item \verb|cll.pl|: for each fold, the list of examples orderd by their probability of being true
\item \verb|areas.csv|: the areas under the Precision-Recall curve and the Receiver Operating Characteristic curve
\item \verb|curve_roc.m|: a Matlab file for plotting the Receiver Operating Characteristic curve
\item \verb|curve_pr.m|: a Matlab file for plotting the Precision-Recall curve
\end{itemize}
\subsection{Learning Examples}
The subfolders \verb|em|, \verb|rib|, \verb|slipcase| and \verb|slipcover| of the \verb|packages/cplint| folder in Yap git distribution

View File

@ -431,8 +431,8 @@ x=[x 1.0]
y=[y 0.0]
k=convhull(x,y)
plot(x(k),y(k),'r-',x,y,'--b+')
A = polyarea(x,y)~n~n
save area_roc.csv A -ascii -append
%A = polyarea(x,y)~n~n
%save area_roc.csv A -ascii -append
",
[]).
@ -462,8 +462,8 @@ set(gca,'YLim',[0.0 1.0])
rec=[0.0 rec 1.0];
prec=[0.0 prec 0.0];
plot(rec,prec,'--*k')
A=polyarea(rec,prec)
save area_pr.csv A -ascii -append
%A=polyarea(rec,prec)
%save area_pr.csv A -ascii -append
~n~n",
[]).