| Safe Haskell | None |
|---|---|
| Language | Haskell2010 |
LightGBM.Parameters
Contents
Description
Parameter types for LightGBM.
Parameter details are documented in the LightGBM documentation.
Note that some of the parameters listed in the documentation are
not exposed here since they're set implicitly through other parts
of the API. For instance, the task param is set in the
LightGBM.Model API, and the header param is set in the
LightGBM.DataSet API.
- data Param
- = Objective Application
- | BoostingType Booster
- | TrainingData FilePath
- | ValidationData [FilePath]
- | PredictionData FilePath
- | Iterations Natural
- | LearningRate PositiveDouble
- | NumLeaves PositiveInt
- | Parallelism ParallelismStyle
- | NumThreads Natural
- | Device Device
- | RandomSeed Int
- | MaxDepth Natural
- | MinDataInLeaf Natural
- | MinSumHessianInLeaf NonNegativeDouble
- | BaggingFraction LeftOpenProperFraction
- | BaggingFreq PositiveInt
- | BaggingFractionSeed Int
- | FeatureFraction LeftOpenProperFraction
- | FeatureFractionSeed Int
- | EarlyStoppingRounds PositiveInt
- | Regularization_L1 NonNegativeDouble
- | Regularization_L2 NonNegativeDouble
- | MaxDeltaStep PositiveDouble
- | MinSplitGain NonNegativeDouble
- | MinDataPerGroup PositiveInt
- | MaxCatThreshold PositiveInt
- | CatSmooth NonNegativeDouble
- | CatL2 NonNegativeDouble
- | MaxCatToOneHot PositiveInt
- | TopK PositiveInt
- | MonotoneConstraint [Direction]
- | MaxBin IntGreaterThanOne
- | MinDataInBin PositiveInt
- | DataRandomSeed Int
- | OutputModel FilePath
- | InputModel FilePath
- | OutputResult FilePath
- | PrePartition Bool
- | IsSparse Bool
- | TwoRoundLoading Bool
- | SaveBinary Bool
- | Verbosity VerbosityLevel
- | LabelColumn (ColumnSelector Natural)
- | WeightColumn (ColumnSelector Natural)
- | QueryColumn (ColumnSelector Natural)
- | IgnoreColumns [ColumnSelector Natural]
- | CategoricalFeatures [ColumnSelector Int32]
- | BinConstructSampleCount PositiveInt
- | UseMissing Bool
- | ZeroAsMissing Bool
- | InitScoreFile FilePath
- | ValidInitScoreFile [FilePath]
- | ForcedSplits FilePath
- | Sigmoid PositiveDouble
- | Alpha OpenProperFraction
- | BoostFromAverage Bool
- | RegSqrt Bool
- | Metric [Metric]
- | MetricFreq PositiveInt
- | TrainingMetric Bool
- data PredictionParam
- data Application
- data BinaryClassParam
- data Booster
- data DARTParam
- data Device
- data Direction
- data FairRegressionParam = FairC PositiveDouble
- data GOSSParam
- data GPUParam
- data LambdaRankParam
- type LocalListenPort = Natural
- type MachineListFile = FilePath
- data Metric
- = MeanAbsoluteError
- | MeanSquareError
- | L2_root
- | QuantileRegression
- | MAPELoss
- | HuberLoss
- | FairLoss
- | PoissonNegLogLikelihood
- | GammaNegLogLikelihood
- | GammaDeviance
- | TweedieNegLogLiklihood
- | NDCG (Maybe NDCGEvalPositions)
- | MAP
- | AUC
- | BinaryLogloss
- | BinaryError
- | MultiLogloss
- | MultiError
- | Xentropy
- | XentLambda
- | KullbackLeibler
- type Minutes = Natural
- data MultiClassStyle
- type NDCGEvalPositions = [Natural]
- type NumClasses = Natural
- type NumMachines = PositiveInt
- data ParallelismParams
- data ParallelismStyle
- data PoissonRegressionParam = PoissonMaxDeltaStep PositiveDouble
- data RegressionApp
- = L1
- | L2
- | Huber
- | Fair [FairRegressionParam]
- | Poisson [PoissonRegressionParam]
- | Quantile
- | MAPE
- | Gamma
- | Tweedie [TweedieRegressionParam]
- data TweedieRegressionParam = TweedieVariancePower OneToTwoLeftSemiClosed
- data VerbosityLevel
- data XEApp
- data ColumnSelector a
- colSelArgument :: Show a => ColumnSelector a -> String
Parameters
Parameters control the behavior of lightGBM.
Constructors
data PredictionParam Source #
Constructors
| PredictRawScore Bool | True = raw scores only, False = transformed scores |
| PredictLeafIndex Bool | True = predict with leaf index |
| PredictContrib Bool | True = estimate how each feature contributes to the prediction |
| NumIterationsPredict Natural | how many trained iterations are used in prediction |
| PredEarlyStop Bool | True = use early stopping on the prediction (may degrade accuracy) |
| PredEarlyStopFreq Natural | |
| PredEarlyStopMargin Double |
Instances
data Application Source #
LightGBM can be used for a variety of applications
Constructors
| Regression RegressionApp | |
| BinaryClassification [BinaryClassParam] | |
| MultiClass MultiClassStyle NumClasses | |
| CrossEntropy XEApp | |
| LambdaRank [LambdaRankParam] | A ranking algorithm |
Instances
data BinaryClassParam Source #
Constructors
| IsUnbalance Bool | Set to true if training data are unbalanced |
| ScalePosWeight PositiveDouble |
Different types of Boosting approaches
Parameters exclusively for the DART booster
Constructors
| DropRate ProperFraction | Dropout rate |
| SkipDrop ProperFraction | Probablility of skipping a drop |
| MaxDrop PositiveInt | Max number of dropped trees on one iteration |
| UniformDrop Bool | |
| XGBoostDARTMode Bool | |
| DropSeed Int |
Constructors
| Increasing | |
| Decreasing | |
| NoConstraint |
data FairRegressionParam Source #
Constructors
| FairC PositiveDouble |
Constructors
| TopRate ProperFraction | |
| OtherRate ProperFraction |
Constructors
| GpuPlatformId Natural | |
| GpuDeviceId Natural | |
| GpuUseDP Bool |
data LambdaRankParam Source #
Constructors
| MaxPosition PositiveInt | |
| LabelGain [Double] | Used in LambdaRank |
Instances
type LocalListenPort = Natural Source #
type MachineListFile = FilePath Source #
Constructors
data MultiClassStyle Source #
Multi-classification styles
Constructors
| MultiClassSimple | |
| MultiClassOneVsAll |
Instances
type NDCGEvalPositions = [Natural] Source #
type NumClasses = Natural Source #
type NumMachines = PositiveInt Source #
data ParallelismParams Source #
data ParallelismStyle Source #
data PoissonRegressionParam Source #
Parameters for Poisson regression
Constructors
| PoissonMaxDeltaStep PositiveDouble |
data RegressionApp Source #
Different types of regression metrics
Constructors
| L1 | Absolute error metric |
| L2 | RMS errror metric |
| Huber | |
| Fair [FairRegressionParam] | |
| Poisson [PoissonRegressionParam] | |
| Quantile | |
| MAPE | |
| Gamma | |
| Tweedie [TweedieRegressionParam] |
Instances
data TweedieRegressionParam Source #
Constructors
| TweedieVariancePower OneToTwoLeftSemiClosed | Control Tweedie variance in range [1, 2) - 1 is like Poisson, 2 is like Gamma |
data VerbosityLevel Source #
Instances
Constructors
| XEntropy | |
| XEntropyLambda |
Utilities
data ColumnSelector a Source #
Some parameters are based on column selection either by index or
by name. A ColumnSelector encapsulates this flexibility.
Instances
| Eq a => Eq (ColumnSelector a) Source # | |
| Show a => Show (ColumnSelector a) Source # | |
colSelArgument :: Show a => ColumnSelector a -> String Source #