blog
Class SamplingEngine
java.lang.Object
blog.InferenceEngine
blog.SamplingEngine
public class SamplingEngine
- extends InferenceEngine
Inference engine that answers queries by sampling partial worlds
(possibly with associated weights) from some sampling distribution,
and collecting statistics over these samples.
The SamplingEngine constructor looks for the following properties
in the properties table that is passed in:
- samplerClass
- Name of the Sampler subclass to use.
Default is blog.LWSampler.
- numSamples
- Number of samples that
answerQueries
will run.
Default is 10,000.
- burnIn
- Number of initial samples to treat as a burn-in period. Samples
during this period will not be used to compute answers to queries.
Default is 0.
- reportInterval
- Number of samples between progress reports. A progress report
just involves printing how many samples have been done and
how long the sampler has been running in this trial.
Default is 500.
The property list is also passed to the sampler's constructor.
Constructor Summary |
SamplingEngine(Model model,
java.util.Properties properties)
Creates a new sampling engine for the given BLOG model, with
configuration parameters specified by the given properties table. |
Method Summary |
void |
answerQueries()
Computes the answers to the specified queries given the specified
evidence. |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
SamplingEngine
public SamplingEngine(Model model,
java.util.Properties properties)
- Creates a new sampling engine for the given BLOG model, with
configuration parameters specified by the given properties table.
answerQueries
public void answerQueries()
- Description copied from class:
InferenceEngine
- Computes the answers to the specified queries given the specified
evidence. Records the answers by calling the appropriate methods
(e.g.,
updateStats
) on the specified Query objects.
- Specified by:
answerQueries
in class InferenceEngine