python-pebl-doc 1.0.2-2 / usr / share / doc / python-pebl / html / evaluator.html

This file is indexed.

/usr/share/doc/python-pebl/html/evaluator.html is in python-pebl-doc 1.0.2-2.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below. 

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
    
    <title>evaluator – Network evaluators &mdash; Pebl v1.0.1 documentation</title>
    <link rel="stylesheet" href="_static/default.css" type="text/css" />
    <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
    <script type="text/javascript">
      var DOCUMENTATION_OPTIONS = {
        URL_ROOT:    '',
        VERSION:     '1.0.1',
        COLLAPSE_INDEX: false,
        FILE_SUFFIX: '.html',
        HAS_SOURCE:  true
      };
    </script>
    <script type="text/javascript" src="_static/jquery.js"></script>
    <script type="text/javascript" src="_static/underscore.js"></script>
    <script type="text/javascript" src="_static/doctools.js"></script>
    <link rel="top" title="Pebl v1.0.1 documentation" href="index.html" />
    <link rel="up" title="API Reference" href="apiref.html" />
    <link rel="next" title="learner.exhaustive – Exhaustive learners" href="learner/exhaustive.html" />
    <link rel="prev" title="discretizer – Discretization algorithms" href="discretizer.html" /> 
  </head>
  <body>
    <div class="related">
      <h3>Navigation</h3>
      <ul>
        <li class="right" style="margin-right: 10px">
          <a href="genindex.html" title="General Index"
             accesskey="I">index</a></li>
        <li class="right" >
          <a href="py-modindex.html" title="Python Module Index"
             >modules</a> |</li>
        <li class="right" >
          <a href="learner/exhaustive.html" title="learner.exhaustive – Exhaustive learners"
             accesskey="N">next</a> |</li>
        <li class="right" >
          <a href="discretizer.html" title="discretizer – Discretization algorithms"
             accesskey="P">previous</a> |</li>
        <li><a href="index.html">Pebl v1.0.1 documentation</a> &raquo;</li>
          <li><a href="apiref.html" accesskey="U">API Reference</a> &raquo;</li> 
      </ul>
    </div>  

    <div class="document">
      <div class="documentwrapper">
        <div class="bodywrapper">
          <div class="body">
            
  <div class="section" id="module-pebl.evaluator">
<span id="evaluator-network-evaluators"></span><h1><tt class="xref py py-mod docutils literal"><span class="pre">evaluator</span></tt> &#8211; Network evaluators<a class="headerlink" href="#module-pebl.evaluator" title="Permalink to this headline"></a></h1>
<p>Greedy learning algorithms work by scoring small, local changes to existing
networks. To do this efficiently, one must maintain state to eliminate
redundant computation or unnecessary cache retrievals &#8211; that is, we should
only rescore nodes that have changed.  Maintaining this state can make
implementing efficient versions of learning algorithms more difficult than a
naive implementation.</p>
<p>The classes in this module provide helpers that encapsulate all the
state-management complexities required for efficient scoring.  As long as
callers make changes to networks in a transactional manner (using the provided
methods), networks will be scored efficiently without redendant computation.</p>
<p>The main evaluator is the SmartNetworkEvaluator class. It&#8217;s interface is described below.</p>
<p>Note: Most user&#8217;s shouldn&#8217;t need to use this module directly. All included
learners encapsulate this functionality. This is really only for writing custom
learners.</p>
<div class="section" id="localscorecache">
<h2>LocalscoreCache<a class="headerlink" href="#localscorecache" title="Permalink to this headline"></a></h2>
<p>Although most users will never use the localscore cache directly, using pebl
with large datasets will require setting the maximum size of the cache to avoid
memory issues. There is only one relevant configuration parameter.</p>
<dl class="confparam">
<dt id="confparam-localscore_cache.maxsize">
<tt class="descname">localscore_cache.maxsize</tt><a class="headerlink" href="#confparam-localscore_cache.maxsize" title="Permalink to this definition"></a></dt>
<dd><p>Max number of localscores to cache. Default=-1 means unlimited size.
default=-1</p>
</dd></dl>

</div>
<div class="section" id="smartnetworkevaluator">
<h2>SmartNetworkEvaluator<a class="headerlink" href="#smartnetworkevaluator" title="Permalink to this headline"></a></h2>
<dl class="class">
<dt id="pebl.evaluator.SmartNetworkEvaluator">
<em class="property">class </em><tt class="descclassname">pebl.evaluator.</tt><tt class="descname">SmartNetworkEvaluator</tt><big>(</big><em>data_</em>, <em>network_</em>, <em>prior_=None</em>, <em>localscore_cache=None</em><big>)</big><a class="headerlink" href="#pebl.evaluator.SmartNetworkEvaluator" title="Permalink to this definition"></a></dt>
<dd><p>Create a &#8216;smart&#8217; network evaluator.</p>
<p>This network evaluator eliminates redundant computation by keeping
track of changes to network and only rescoring the changes. This
requires that all changes to the network are done through this
evaluator&#8217;s methods.</p>
<dl class="docutils">
<dt>The network can be altered by the following methods:</dt>
<dd><ul class="first last simple">
<li>alter_network</li>
<li>score_network</li>
<li>randomize_network</li>
<li>clear_network</li>
</ul>
</dd>
</dl>
<p>The last change applied can be &#8216;undone&#8217; with restore_network</p>
<dl class="method">
<dt id="pebl.evaluator.SmartNetworkEvaluator.alter_network">
<tt class="descname">alter_network</tt><big>(</big><em>add=</em><span class="optional">[</span><span class="optional">]</span>, <em>remove=</em><span class="optional">[</span><span class="optional">]</span><big>)</big><a class="headerlink" href="#pebl.evaluator.SmartNetworkEvaluator.alter_network" title="Permalink to this definition"></a></dt>
<dd><p>Alter the network while retaining the ability to <em>quickly</em> undo the changes.</p>
</dd></dl>

<dl class="method">
<dt id="pebl.evaluator.SmartNetworkEvaluator.clear_network">
<tt class="descname">clear_network</tt><big>(</big><big>)</big><a class="headerlink" href="#pebl.evaluator.SmartNetworkEvaluator.clear_network" title="Permalink to this definition"></a></dt>
<dd><p>Clear all edges from the network.</p>
</dd></dl>

<dl class="method">
<dt id="pebl.evaluator.SmartNetworkEvaluator.randomize_network">
<tt class="descname">randomize_network</tt><big>(</big><big>)</big><a class="headerlink" href="#pebl.evaluator.SmartNetworkEvaluator.randomize_network" title="Permalink to this definition"></a></dt>
<dd><p>Randomize the network edges.</p>
</dd></dl>

<dl class="method">
<dt id="pebl.evaluator.SmartNetworkEvaluator.restore_network">
<tt class="descname">restore_network</tt><big>(</big><big>)</big><a class="headerlink" href="#pebl.evaluator.SmartNetworkEvaluator.restore_network" title="Permalink to this definition"></a></dt>
<dd><p>Undo the last change to the network (and score).</p>
<dl class="docutils">
<dt>Undo the last change performed by any of these methods:</dt>
<dd><ul class="first last simple">
<li>score_network</li>
<li>alter_network</li>
<li>randomize_network</li>
<li>clear_network</li>
</ul>
</dd>
</dl>
</dd></dl>

<dl class="method">
<dt id="pebl.evaluator.SmartNetworkEvaluator.score_network">
<tt class="descname">score_network</tt><big>(</big><em>net=None</em><big>)</big><a class="headerlink" href="#pebl.evaluator.SmartNetworkEvaluator.score_network" title="Permalink to this definition"></a></dt>
<dd><p>Score a network.</p>
<p>If net is provided, scores that. Otherwise, score network previously
set.</p>
</dd></dl>

</dd></dl>

</div>
<div class="section" id="network-evaluators-for-use-with-missing-values">
<h2>Network Evaluators for use with Missing Values<a class="headerlink" href="#network-evaluators-for-use-with-missing-values" title="Permalink to this headline"></a></h2>
<p>Scoring networks with missing values requires use of sampling algorithms to
sample over the space of possible completions for the missing values. Pebl
provides a few algorithms for this.</p>
<div class="section" id="configuration-parameters">
<h3>Configuration Parameters<a class="headerlink" href="#configuration-parameters" title="Permalink to this headline"></a></h3>
<dl class="confparam">
<dt id="confparam-evaluator.missingdata_evaluator">
<tt class="descname">evaluator.missingdata_evaluator</tt><a class="headerlink" href="#confparam-evaluator.missingdata_evaluator" title="Permalink to this definition"></a></dt>
<dd><dl class="docutils">
<dt>Evaluator to use for handling missing data. Choices include:</dt>
<dd><ul class="first">
<li><p class="first">gibbs: Gibb&#8217;s sampling</p>
</li>
<li><p class="first">maxentropy_gibbs: Gibbs&#8217;s sampling over all completions of the
missing values that result in maximum entropy discretization for the
variables.</p>
</li>
<li><dl class="first docutils">
<dt>exact: exact enumeration of all possible missing values (only</dt>
<dd><p class="first last">useable when there are few missing values)</p>
</dd>
</dl>
</li>
</ul>
<p class="last">default=gibbs</p>
</dd>
</dl>
</dd></dl>

<dl class="confparam">
<dt id="confparam-gibbs.burnin">
<tt class="descname">gibbs.burnin</tt><a class="headerlink" href="#confparam-gibbs.burnin" title="Permalink to this definition"></a></dt>
<dd><p>Burn-in period for the gibbs sampler (specified as a multiple of the number of missing values)
default=10</p>
</dd></dl>

<dl class="confparam">
<dt id="confparam-gibbs.stopping_criteria">
<tt class="descname">gibbs.stopping_criteria</tt><a class="headerlink" href="#confparam-gibbs.stopping_criteria" title="Permalink to this definition"></a></dt>
<dd><p>Stopping criteria for the gibbs sampler.</p>
<p>Should be a valid python expression that evaluates to true when gibbs
should stop. It can use the following variables:</p>
<blockquote>
<div><ul class="simple">
<li>iters: number of iterations</li>
<li>n: number of missing values</li>
</ul>
</div></blockquote>
<p>Examples:</p>
<blockquote>
<div><ul class="simple">
<li>iters &gt; n**2  (for n-squared iterations)</li>
<li>iters &gt; 100   (for 100 iterations)</li>
</ul>
</div></blockquote>
<p>default=iters &gt; n**2</p>
</dd></dl>

</div>
<div class="section" id="missingdatanetworkevaluator">
<h3>MissingDataNetworkEvaluator<a class="headerlink" href="#missingdatanetworkevaluator" title="Permalink to this headline"></a></h3>
<dl class="class">
<dt id="pebl.evaluator.MissingDataNetworkEvaluator">
<em class="property">class </em><tt class="descclassname">pebl.evaluator.</tt><tt class="descname">MissingDataNetworkEvaluator</tt><big>(</big><em>data_</em>, <em>network_</em>, <em>prior_=None</em>, <em>localscore_cache=None</em>, <em>**options</em><big>)</big><a class="headerlink" href="#pebl.evaluator.MissingDataNetworkEvaluator" title="Permalink to this definition"></a></dt>
<dd><p>Create a network evaluator for use with missing values.</p>
<p>This evaluator uses a Gibb&#8217;s sampler for sampling over the space of
possible completions for the missing values.</p>
<p>For more information about Gibb&#8217;s sampling, consult:</p>
<blockquote>
<div><ol class="arabic simple">
<li><a class="reference external" href="http://en.wikipedia.org/wiki/Gibbs_sampling">http://en.wikipedia.org/wiki/Gibbs_sampling</a></li>
<li>D. Heckerman. A Tutorial on Learning with Bayesian Networks. 
Microsoft Technical Report MSR-TR-95-06, 1995. p.21-22.</li>
</ol>
</div></blockquote>
<p>Any config param for &#8216;gibbs&#8217; can be passed in via options.
Use just the option part of the parameter name.</p>
<dl class="method">
<dt id="pebl.evaluator.MissingDataNetworkEvaluator.score_network">
<tt class="descname">score_network</tt><big>(</big><em>net=None</em>, <em>gibbs_state=None</em><big>)</big><a class="headerlink" href="#pebl.evaluator.MissingDataNetworkEvaluator.score_network" title="Permalink to this definition"></a></dt>
<dd><p>Score a network.</p>
<p>If net is provided, scores that. Otherwise, score network previously
set.</p>
<p>The default stopping criteria is to run for n**2 iterations.</p>
<p>gibbs_state is the state of a previous run of the Gibb&#8217;s sampler.  With
this, one can do the following:</p>
<div class="highlight-python"><div class="highlight"><pre><span class="n">myeval</span> <span class="o">=</span> <span class="n">evaluator</span><span class="o">.</span><span class="n">MissingDataNetworkEvaluator</span><span class="p">(</span><span class="o">...</span><span class="p">)</span>
<span class="n">myeval</span><span class="o">.</span><span class="n">score_network</span><span class="p">(</span><span class="o">...</span><span class="p">)</span>
<span class="n">gibbs_state</span> <span class="o">=</span> <span class="n">myeval</span><span class="o">.</span><span class="n">gibbs_state</span>
<span class="n">cPickle</span><span class="o">.</span><span class="n">dump</span><span class="p">(</span><span class="n">gibbs_state</span><span class="p">,</span> <span class="s">&#39;gibbs_state.txt&#39;</span><span class="p">)</span>

<span class="c"># look at results, do other analysis, etc</span>
<span class="c"># If we decide that we need further Gibb&#39;s sampler iterations, we</span>
<span class="c"># don&#39;t need to restart</span>
<span class="n">gibbs_state</span> <span class="o">=</span> <span class="n">cPickle</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="nb">open</span><span class="p">(</span><span class="s">&#39;gibbs_state.txt&#39;</span><span class="p">))</span>
<span class="n">myeval</span> <span class="o">=</span> <span class="n">evaluator</span><span class="o">.</span><span class="n">MissingDataNetworkEvaluator</span><span class="p">(</span><span class="o">...</span><span class="p">)</span>

<span class="c"># continue with the previous run of the Gibb&#39;s sampler</span>
<span class="n">myeval</span><span class="o">.</span><span class="n">score_network</span><span class="p">(</span>
    <span class="n">gibbs_state</span><span class="o">=</span><span class="n">gibbs_state</span><span class="p">,</span>
    <span class="n">stopping_criteria</span><span class="o">=</span><span class="k">lambda</span> <span class="n">i</span><span class="p">,</span><span class="n">N</span><span class="p">:</span> <span class="n">i</span><span class="o">&gt;</span><span class="mi">200</span><span class="o">*</span><span class="n">N</span><span class="o">**</span><span class="mi">2</span>
<span class="p">)</span>
</pre></div>
</div>
</dd></dl>

</dd></dl>

<dl class="class">
<dt id="pebl.evaluator.MissingDataMaximumEntropyNetworkEvaluator">
<em class="property">class </em><tt class="descclassname">pebl.evaluator.</tt><tt class="descname">MissingDataMaximumEntropyNetworkEvaluator</tt><big>(</big><em>data_</em>, <em>network_</em>, <em>prior_=None</em>, <em>localscore_cache=None</em>, <em>**options</em><big>)</big><a class="headerlink" href="#pebl.evaluator.MissingDataMaximumEntropyNetworkEvaluator" title="Permalink to this definition"></a></dt>
<dd><p>MissingDataNetworkEvaluator that uses a different space of completions.</p>
<p>This evaluator only samples from missing value completions that result in a
maximum entropy discretization for the variables with missing values. This
is useful when the rest of the variables are maximum-entropy discretized
because then all variables have the same entropy.</p>
<p>Create a network evaluator for use with missing values.</p>
<p>This evaluator uses a Gibb&#8217;s sampler for sampling over the space of
possible completions for the missing values.</p>
<p>For more information about Gibb&#8217;s sampling, consult:</p>
<blockquote>
<div><ol class="arabic simple">
<li><a class="reference external" href="http://en.wikipedia.org/wiki/Gibbs_sampling">http://en.wikipedia.org/wiki/Gibbs_sampling</a></li>
<li>D. Heckerman. A Tutorial on Learning with Bayesian Networks. 
Microsoft Technical Report MSR-TR-95-06, 1995. p.21-22.</li>
</ol>
</div></blockquote>
<p>Any config param for &#8216;gibbs&#8217; can be passed in via options.
Use just the option part of the parameter name.</p>
</dd></dl>

<dl class="class">
<dt id="pebl.evaluator.MissingDataExactNetworkEvaluator">
<em class="property">class </em><tt class="descclassname">pebl.evaluator.</tt><tt class="descname">MissingDataExactNetworkEvaluator</tt><big>(</big><em>data_</em>, <em>network_</em>, <em>prior_=None</em>, <em>localscore_cache=None</em>, <em>**options</em><big>)</big><a class="headerlink" href="#pebl.evaluator.MissingDataExactNetworkEvaluator" title="Permalink to this definition"></a></dt>
<dd><p>MissingDataNEtworkEvaluator that does an exact enumeration.</p>
<p>This network evaluator enumerates over all possible completions of the
missing values.  Since this is a combinatorial space, this class is only
feasible with datasets with few missing values.</p>
<p>Create a network evaluator for use with missing values.</p>
<p>This evaluator uses a Gibb&#8217;s sampler for sampling over the space of
possible completions for the missing values.</p>
<p>For more information about Gibb&#8217;s sampling, consult:</p>
<blockquote>
<div><ol class="arabic simple">
<li><a class="reference external" href="http://en.wikipedia.org/wiki/Gibbs_sampling">http://en.wikipedia.org/wiki/Gibbs_sampling</a></li>
<li>D. Heckerman. A Tutorial on Learning with Bayesian Networks. 
Microsoft Technical Report MSR-TR-95-06, 1995. p.21-22.</li>
</ol>
</div></blockquote>
<p>Any config param for &#8216;gibbs&#8217; can be passed in via options.
Use just the option part of the parameter name.</p>
</dd></dl>

</div>
</div>
<div class="section" id="factory-functions">
<h2>Factory Functions<a class="headerlink" href="#factory-functions" title="Permalink to this headline"></a></h2>
<dl class="function">
<dt id="pebl.evaluator.fromconfig">
<tt class="descclassname">pebl.evaluator.</tt><tt class="descname">fromconfig</tt><big>(</big><em>data_=None</em>, <em>network_=None</em>, <em>prior_=None</em><big>)</big><a class="headerlink" href="#pebl.evaluator.fromconfig" title="Permalink to this definition"></a></dt>
<dd><p>Create an evaluator based on configuration parameters.</p>
<p>This function will return the correct evaluator based on the relevant
configuration parameters.</p>
</dd></dl>

</div>
</div>


          </div>
        </div>
      </div>
      <div class="sphinxsidebar">
        <div class="sphinxsidebarwrapper">
  <h3><a href="index.html">Table Of Contents</a></h3>
  <ul>
<li><a class="reference internal" href="#"><tt class="docutils literal"><span class="pre">evaluator</span></tt> &#8211; Network evaluators</a><ul>
<li><a class="reference internal" href="#localscorecache">LocalscoreCache</a></li>
<li><a class="reference internal" href="#smartnetworkevaluator">SmartNetworkEvaluator</a></li>
<li><a class="reference internal" href="#network-evaluators-for-use-with-missing-values">Network Evaluators for use with Missing Values</a><ul>
<li><a class="reference internal" href="#configuration-parameters">Configuration Parameters</a></li>
<li><a class="reference internal" href="#missingdatanetworkevaluator">MissingDataNetworkEvaluator</a></li>
</ul>
</li>
<li><a class="reference internal" href="#factory-functions">Factory Functions</a></li>
</ul>
</li>
</ul>

  <h4>Previous topic</h4>
  <p class="topless"><a href="discretizer.html"
                        title="previous chapter"><tt class="docutils literal docutils literal docutils literal"><span class="pre">discretizer</span></tt> &#8211; Discretization algorithms</a></p>
  <h4>Next topic</h4>
  <p class="topless"><a href="learner/exhaustive.html"
                        title="next chapter"><tt class="docutils literal"><span class="pre">learner.exhaustive</span></tt> &#8211; Exhaustive learners</a></p>
  <h3>This Page</h3>
  <ul class="this-page-menu">
    <li><a href="_sources/evaluator.txt"
           rel="nofollow">Show Source</a></li>
  </ul>
<div id="searchbox" style="display: none">
  <h3>Quick search</h3>
    <form class="search" action="search.html" method="get">
      <input type="text" name="q" size="18" />
      <input type="submit" value="Go" />
      <input type="hidden" name="check_keywords" value="yes" />
      <input type="hidden" name="area" value="default" />
    </form>
    <p class="searchtip" style="font-size: 90%">
    Enter search terms or a module, class or function name.
    </p>
</div>
<script type="text/javascript">$('#searchbox').show(0);</script>
        </div>
      </div>
      <div class="clearer"></div>
    </div>
    <div class="related">
      <h3>Navigation</h3>
      <ul>
        <li class="right" style="margin-right: 10px">
          <a href="genindex.html" title="General Index"
             >index</a></li>
        <li class="right" >
          <a href="py-modindex.html" title="Python Module Index"
             >modules</a> |</li>
        <li class="right" >
          <a href="learner/exhaustive.html" title="learner.exhaustive – Exhaustive learners"
             >next</a> |</li>
        <li class="right" >
          <a href="discretizer.html" title="discretizer – Discretization algorithms"
             >previous</a> |</li>
        <li><a href="index.html">Pebl v1.0.1 documentation</a> &raquo;</li>
          <li><a href="apiref.html" >API Reference</a> &raquo;</li> 
      </ul>
    </div>
    <div class="footer">
        &copy; Copyright 2008, Abhik Shah.
      Last updated on Apr 29, 2011.
      Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 1.0.7.
    </div>
  </body>
</html>

APT Browse - Built by Thomas Orozco.