calhoun.analysis.crf.solver
Class TwoPassOptimizer

java.lang.Object
  calhoun.analysis.crf.solver.TwoPassOptimizer

All Implemented Interfaces:

CRFTraining

public class TwoPassOptimizer
extends java.lang.Object
implements CRFTraining

a two pass optimizer that does an initial optimization and then uses the weights generated from that as the start of a second optimization. It is configured with two separate CRFOptimizer oibjects for the two passes and additionally needs a CRFObjectiveFunctionGradient for the first pass.

This optimizer has two configuration properties that must be set and control the optimization process

• firstPass - A reference to a configured CRFTraining object that will be used to optimize the weights in the first pass of the optimization.
• secondPass - A reference to a configured CRFTraining object that will be used to optimize the weights in the second pass of the optimization. The second pass will begin using the weights computed during the first pass.

  Constructor Summary

  TwoPassOptimizer()
  

  Method Summary

  CRFTraining	getFirstPass()
  CRFTraining	getSecondPass()
  double[]	optimize(ModelManager fm, java.util.List<? extends TrainingSequence<?>> data) find the set of weights which maximizes the value of the objective function.
  void	setFirstPass(CRFTraining firstPass)
  void	setSecondPass(CRFTraining secondPass)
  void	setStarts(double[] weights) Sets the starting weights for the optimization.

  Methods inherited from class java.lang.Object

  equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

  Constructor Detail

  TwoPassOptimizer

  public TwoPassOptimizer()

  Method Detail

  optimize

  public double[] optimize(ModelManager fm,
                           java.util.List<? extends TrainingSequence<?>> data)

  Description copied from interface: CRFTraining

  find the set of weights which maximizes the value of the objective function. The CRFObjectiveFunctionGradient object will already be configured witht eh appropriate training data, so it appears here as a pure function evaluation.

  Specified by:

  optimize in interface CRFTraining

  Parameters:

  fm - the model to train on.

  data - the training data to use for training.

  Returns:

  an array of doubles containing the optimal weights.

  setStarts

  public void setStarts(double[] weights)

  Description copied from interface: CRFTraining

  Sets the starting weights for the optimization.

  Specified by:

  setStarts in interface CRFTraining

  Parameters:

  weights - an array of weights to use as the starting point for the optimizer. The optimzier is not required to use this as a starting point.

  getFirstPass

  public CRFTraining getFirstPass()

  Returns:

  the firstPass

  setFirstPass

  public void setFirstPass(CRFTraining firstPass)

  Parameters:

  firstPass - the firstPass to set

  getSecondPass

  public CRFTraining getSecondPass()

  Returns:

  the secondPass

  setSecondPass

  public void setSecondPass(CRFTraining secondPass)

  Parameters:

  secondPass - the secondPass to set