LightGBM Namespace Reference

desc and descl2 fields must be written in reStructuredText format More...

Data Structures
class	Application
	The main entrance of LightGBM. this application has two tasks: Train and Predict. Train task will train a new model Predict task will predict the scores of test data using existing model, and save the score to disk. More...
class	ArrayArgs
	Contains some operation for a array, e.g. ArgMax, TopK. More...
class	AUCMetric
	Auc Metric for binary classification task. More...
class	Bin
	Interface for bin data. This class will store bin data for one feature. unlike OrderedBin, this class will store data by original order. Note that it may cause cache misses when construct histogram, but it doesn't need to re-order operation, So it will be faster than OrderedBin for dense feature. More...
class	BinaryErrorMetric
	Error rate metric for binary classification task. More...
class	BinaryLogloss
	Objective function for binary classification. More...
class	BinaryLoglossMetric
	Log loss metric for binary classification task. More...
class	BinaryMetric
	Metric for binary classification task. Use static class "PointWiseLossCalculator" to calculate loss point-wise. More...
class	BinIterator
	Iterator for one bin column. More...
class	BinMapper
	This class used to convert feature values into bin, and store some meta information for bin. More...
class	Booster
class	Boosting
	The interface for Boosting. More...
class	BruckMap
	The network structure for all_gather. More...
struct	Config
class	CrossEntropy
	Objective function for cross-entropy (with optional linear weights) More...
class	CrossEntropyLambda
	Objective function for alternative parameterization of cross-entropy (see top of file for explanation) More...
class	CrossEntropyLambdaMetric
class	CrossEntropyMetric
class	CSVParser
class	DART
	DART algorithm implementation. including Training, prediction, bagging. More...
class	DataParallelTreeLearner
	Data parallel learning algorithm. Workers use local data to construct histograms locally, then sync up global histograms. It is recommonded used when data is large or #feature is small. More...
class	DataPartition
	DataPartition is used to store the the partition of data on tree. More...
class	Dataset
	The main class of data set, which are used to traning or validation. More...
class	DatasetLoader
class	DCGCalculator
	Static class, used to calculate DCG score. More...
class	Dense4bitsBin
class	Dense4bitsBinIterator
class	DenseBin
	Used to store bins for dense feature Use template to reduce memory cost. More...
class	DenseBinIterator
class	FairLossMetric
	Fair loss for regression task. More...
class	FeatureGroup
	Using to store data and providing some operations on one feature group. More...
class	FeatureHistogram
	FeatureHistogram is used to construct and store a histogram for a feature. More...
class	FeatureMetainfo
class	FeatureParallelTreeLearner
	Feature parallel learning algorithm. Different machine will find best split on different features, then sync global best split It is recommonded used when data is small or #feature is large. More...
class	GammaDevianceMetric
class	GammaMetric
class	GBDT
	GBDT algorithm implementation. including Training, prediction, bagging. More...
class	GBDT_Accessor
class	GBDTBase
class	GOSS
class	GPUTreeLearner
struct	HistogramBinEntry
	Store data for one histogram bin. More...
class	HistogramPool
class	HuberLossMetric
	Huber loss for regression task. More...
class	KullbackLeiblerDivergence
class	L1Metric
	L1 loss for regression task. More...
class	L2Metric
	L2 loss for regression task. More...
class	LambdarankNDCG
	Objective function for Lambdrank with NDCG. More...
class	LeafSplits
	used to find split candidates for a leaf More...
class	LibSVMParser
struct	LightSplitInfo
class	Linkers
	An network basic communication warpper. Will warp low level communication methods, e.g. mpi, socket and so on. This class will wrap all linkers to other machines if needs. More...
struct	LocalFile
class	Log
	A static Log class. More...
class	MAPEMetric
	Mape regression loss for regression task. More...
class	MapMetric
class	MemApp
class	Metadata
	This class is used to store some meta(non-feature) data for training data, e.g. labels, weights, initial scores, qurey level informations. More...
class	Metric
	The interface of metric. Metric is used to calculate metric result. More...
class	MulticlassMetric
	Metric for multiclass task. Use static class "PointWiseLossCalculator" to calculate loss point-wise. More...
class	MulticlassOVA
	Objective function for multiclass classification, use one-vs-all binary objective function. More...
class	MulticlassSoftmax
	Objective function for multiclass classification, use softmax as objective functions. More...
class	MultiErrorMetric
	L2 loss for multiclass task. More...
class	MultiSoftmaxLoglossMetric
	Logloss for multiclass task. More...
class	NDCGMetric
class	Network
	A static class that contains some collective communication algorithm. More...
class	ObjectiveFunction
	The interface of Objective Function. More...
class	OrderedBin
	Interface for ordered bin data. efficient for construct histogram, especially for sparse bin There are 2 advantages by using ordered bin. More...
class	OrderedSparseBin
	Interface for ordered bin data. efficient for construct histogram, especially for sparse bin There are 2 advantages by using ordered bin. More...
struct	ParameterAlias
class	Parser
	Interface for Parser. More...
class	PipelineReader
	A pipeline file reader, use 2 threads, one read block from file, the other process the block. More...
class	PoissonMetric
	Poisson regression loss for regression task. More...
struct	PredictionEarlyStopConfig
struct	PredictionEarlyStopInstance
class	Predictor
	Used to predict data with input model. More...
class	QuantileMetric
	L2 loss for regression task. More...
class	Random
	A wrapper for random generator. More...
class	RecursiveHalvingMap
	Network structure for recursive halving algorithm. More...
class	RegressionFairLoss
class	RegressionGammaLoss
	Objective function for Gamma regression. More...
class	RegressionHuberLoss
	Huber regression loss. More...
class	RegressionL1loss
	L1 regression loss. More...
class	RegressionL2loss
	Objective function for regression. More...
class	RegressionMAPELOSS
	Mape Regression Loss. More...
class	RegressionMetric
	Metric for regression task. Use static class "PointWiseLossCalculator" to calculate loss point-wise. More...
class	RegressionPoissonLoss
	Objective function for Poisson regression. More...
class	RegressionQuantileloss
class	RegressionTweedieLoss
	Objective function for Tweedie regression. More...
class	RF
	Rondom Forest implementation. More...
class	RMSEMetric
	RMSE loss for regression task. More...
class	ScoreUpdater
	Used to store and update score for data. More...
class	SerialTreeLearner
	Used for learning a tree by single machine. More...
class	SparseBin
class	SparseBinIterator
struct	SplitInfo
	Used to store some information for gain split point. More...
class	TextReader
	Read text data from file. More...
class	Threading
class	Tree
	Tree model. More...
class	TreeLearner
	Interface for tree learner. More...
class	TSVParser
class	TweedieMetric
struct	VirtualFileReader
	An interface for reading files into buffers. More...
struct	VirtualFileWriter
	An interface for writing files from buffers. More...
class	VotingParallelTreeLearner
	Voting based data parallel learning algorithm. Like data parallel, but not aggregate histograms for all features. Here using voting to reduce features, and only aggregate histograms for selected features. When data is large and #feature is large, you can use this to have better speed-up. More...

Typedefs
typedef int32_t	data_size_t
	Type of data size, it is better to use signed type.
typedef float	score_t
	Type of score, and gradients.
typedef float	label_t
	Type of metadata, include weight and label.
typedef int32_t	comm_size_t
using	PredictFunction = std::function< void(const std::vector< std::pair< int, double > > &, double *output)>
typedef void(*	ReduceFunction) (const char *input, char *output, int type_size, comm_size_t array_size)
typedef void(*	ReduceScatterFunction) (char *input, comm_size_t input_size, int type_size, const comm_size_t *block_start, const comm_size_t *block_len, int num_block, char *output, comm_size_t output_size, const ReduceFunction &reducer)
typedef void(*	AllgatherFunction) (char *input, comm_size_t input_size, const comm_size_t *block_start, const comm_size_t *block_len, int num_block, char *output, comm_size_t output_size)

Enumerations
enum	BinType { NumericalBin , CategoricalBin }
enum	MissingType { None , Zero , NaN }
enum	TaskType { kTrain , kPredict , kConvertModel , KRefitTree }
	Types of tasks.
enum	RecursiveHalvingNodeType { Normal , GroupLeader , Other }
	node type on recursive halving algorithm When number of machines is not power of 2, need group machines into power of 2 group. And we can let each group has at most 2 machines. if the group only has 1 machine. this machine is the normal node if the group has 2 machines, this group will have two type of nodes, one is the leader. leader will represent this group and communication with others.
enum class	LogLevel : int { Fatal = -1 , Warning = 0 , Info = 1 , Debug = 2 }
enum	DataType { INVALID , CSV , TSV , LIBSVM }

Functions
LIGHTGBM_EXPORT PredictionEarlyStopInstance	CreatePredictionEarlyStopInstance (const std::string &type, const PredictionEarlyStopConfig &config)
	Create an early stopping algorithm of type type, with given round_period and margin threshold.
std::string	GetBoostingTypeFromModelFile (const char *filename)
double	ObtainAutomaticInitialScore (const ObjectiveFunction *fobj, int class_id)
int	LGBM_APIHandleException (const std::exception &ex)
int	LGBM_APIHandleException (const std::string &ex)
bool	NeedFilter (const std::vector< int > &cnt_in_bin, int total_cnt, int filter_cnt, BinType bin_type)
std::vector< double >	GreedyFindBin (const double *distinct_values, const int *counts, int num_distinct_values, int max_bin, size_t total_cnt, int min_data_in_bin)
std::vector< double >	FindBinWithZeroAsOneBin (const double *distinct_values, const int *counts, int num_distinct_values, int max_bin, size_t total_sample_cnt, int min_data_in_bin)
void	GetBoostingType (const std::unordered_map< std::string, std::string > &params, std::string *boosting)
void	GetObjectiveType (const std::unordered_map< std::string, std::string > &params, std::string *objective)
void	GetMetricType (const std::unordered_map< std::string, std::string > &params, std::vector< std::string > *metric)
void	GetTaskType (const std::unordered_map< std::string, std::string > &params, TaskType *task)
void	GetDeviceType (const std::unordered_map< std::string, std::string > &params, std::string *device_type)
void	GetTreeLearnerType (const std::unordered_map< std::string, std::string > &params, std::string *tree_learner)
bool	CheckMultiClassObjective (const std::string &objective)
std::vector< std::vector< int > >	NoGroup (const std::vector< int > &used_features)
int	GetConfilctCount (const std::vector< bool > &mark, const int *indices, int num_indices, int max_cnt)
void	MarkUsed (std::vector< bool > &mark, const int *indices, int num_indices)
std::vector< std::vector< int > >	FindGroups (const std::vector< std::unique_ptr< BinMapper > > &bin_mappers, const std::vector< int > &find_order, int **sample_indices, const int *num_per_col, size_t total_sample_cnt, data_size_t max_error_cnt, data_size_t filter_cnt, data_size_t num_data, bool is_use_gpu)
std::vector< std::vector< int > >	FastFeatureBundling (std::vector< std::unique_ptr< BinMapper > > &bin_mappers, int **sample_indices, const int *num_per_col, size_t total_sample_cnt, const std::vector< int > &used_features, double max_conflict_rate, data_size_t num_data, data_size_t min_data, double sparse_threshold, bool is_enable_sparse, bool is_use_gpu)
void	GetStatistic (const char *str, int *comma_cnt, int *tab_cnt, int *colon_cnt)
int	GetLabelIdxForLibsvm (std::string &str, int num_features, int label_idx)
int	GetLabelIdxForTSV (std::string &str, int num_features, int label_idx)
int	GetLabelIdxForCSV (std::string &str, int num_features, int label_idx)
void	getline (std::stringstream &ss, std::string &line, const VirtualFileReader *reader, std::vector< char > &buffer, size_t buffer_size)
void	SyncUpGlobalBestSplit (char *input_buffer_, char *output_buffer_, SplitInfo *smaller_best_split, SplitInfo *larger_best_split, int max_cat_threshold)
std::function< std::vector< double >(int row_idx)>	RowFunctionFromDenseMatric (const void *data, int num_row, int num_col, int data_type, int is_row_major)
std::function< std::vector< std::pair< int, double > >(int row_idx)>	RowPairFunctionFromDenseMatric (const void *data, int num_row, int num_col, int data_type, int is_row_major)

Variables
const int	kDefaultNumLeaves = 31
const score_t	kMinScore = -std::numeric_limits<score_t>::infinity()
const score_t	kEpsilon = 1e-15f
const double	kZeroThreshold = 1e-35f
const std::string	kModelVersion = "v2"
const std::string	kHdfsProto = "hdfs://"
const size_t	kNumFastIndex = 64

Detailed Description

desc and descl2 fields must be written in reStructuredText format

This file is auto generated by LightGBM\helpers\parameter_generator.py from LightGBM\include\LightGBM\config.h file.