hex.grid

Class HyperSpaceWalker.RandomDiscreteValueWalker<MP extends Model.Parameters>

java.lang.Object

hex.grid.HyperSpaceWalker.BaseWalker<MP,HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria>

hex.grid.HyperSpaceWalker.RandomDiscreteValueWalker<MP>

All Implemented Interfaces:

HyperSpaceWalker<MP,HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria>

Enclosing interface:

HyperSpaceWalker<MP extends Model.Parameters,C extends HyperSpaceSearchCriteria>

public static class HyperSpaceWalker.RandomDiscreteValueWalker<MP extends Model.Parameters>
extends HyperSpaceWalker.BaseWalker<MP,HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria>

Hyperparameter space walker which visits random combinations of hyperparameters whose possible values are given in explicit lists as they are with CartesianWalker.

Nested Class Summary

Nested classes/interfaces inherited from class hex.grid.HyperSpaceWalker.BaseWalker

HyperSpaceWalker.BaseWalker.WalkerFactory<MP extends Model.Parameters,C extends HyperSpaceSearchCriteria>

Nested classes/interfaces inherited from interface hex.grid.HyperSpaceWalker

HyperSpaceWalker.BaseWalker<MP extends Model.Parameters,C extends HyperSpaceSearchCriteria>, HyperSpaceWalker.CartesianWalker<MP extends Model.Parameters>, HyperSpaceWalker.HyperSpaceIterator<MP extends Model.Parameters>, HyperSpaceWalker.RandomDiscreteValueWalker<MP extends Model.Parameters>

Field Summary

Fields inherited from class hex.grid.HyperSpaceWalker.BaseWalker

_hyperParamNames, _hyperParamNamesSubspace, _hyperParams, _hyperParamSubspaces, _maxHyperSpaceSize, _search_criteria, SUBSPACES

Constructor Summary

Constructors

Constructor and Description
RandomDiscreteValueWalker(MP params, java.util.Map<java.lang.String,java.lang.Object[]> hyperParams, ModelParametersBuilderFactory<MP> paramsBuilderFactory, HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria search_criteria)

Method Summary

Modifier and Type	Method and Description
long	estimateGridWork(long maxModels) Return estimated grid work.
HyperSpaceWalker.HyperSpaceIterator<MP>	iterator() Returns an iterator to traverse this hyper-space.
boolean	stopEarly(Model model, ScoringInfo[] sk) Based on the last model, the given array of ScoringInfo, and our stopping criteria should we stop early?

Methods inherited from class hex.grid.HyperSpaceWalker.BaseWalker

computeMaxSizeOfHyperSpace, getAllHyperParamNamesInSubspaces, getHyperParamNames, getHyperParams, getMaxHyperSpaceSize, getModelParams, getParametersBuilderFactory, getParams, hypers, integerHash, mergeHashMaps, search_criteria

Methods inherited from class java.lang.Object

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

Methods inherited from interface hex.grid.HyperSpaceWalker

getAllHyperParamNames

Constructor Detail

RandomDiscreteValueWalker

public RandomDiscreteValueWalker(MP params,
                                 java.util.Map<java.lang.String,java.lang.Object[]> hyperParams,
                                 ModelParametersBuilderFactory<MP> paramsBuilderFactory,
                                 HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria search_criteria)

Method Detail

stopEarly

public boolean stopEarly(Model model,
                         ScoringInfo[] sk)

Based on the last model, the given array of ScoringInfo, and our stopping criteria should we stop early?

Specified by:

stopEarly in interface HyperSpaceWalker<MP extends Model.Parameters,HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria>

Overrides:

stopEarly in class HyperSpaceWalker.BaseWalker<MP extends Model.Parameters,HyperSpaceSearchCriteria.RandomDiscreteValueSearchCriteria>

iterator

public HyperSpaceWalker.HyperSpaceIterator<MP> iterator()

Description copied from interface: HyperSpaceWalker

Returns an iterator to traverse this hyper-space.

Returns:

an iterator

estimateGridWork

public long estimateGridWork(long maxModels)

Description copied from interface: HyperSpaceWalker

Return estimated grid work. Can return Long.MAX_VALUE if no estimate is available.

Returns:

estimate of grid work