LightGBM::VotingParallelTreeLearner< TREELEARNER_T > Class Template Reference

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...

#include <parallel_tree_learner.h>

Inheritance diagram for LightGBM::VotingParallelTreeLearner< TREELEARNER_T >:

Public Member Functions
	VotingParallelTreeLearner (const Config *config)
void	Init (const Dataset *train_data, bool is_constant_hessian) override
void	ResetConfig (const Config *config) override

Protected Member Functions
void	BeforeTrain () override
bool	BeforeFindBestSplit (const Tree *tree, int left_leaf, int right_leaf) override
void	FindBestSplits () override
void	FindBestSplitsFromHistograms (const std::vector< int8_t > &is_feature_used, bool use_subtract) override
void	Split (Tree *tree, int best_Leaf, int *left_leaf, int *right_leaf) override
data_size_t	GetGlobalDataCountInLeaf (int leaf_idx) const override
void	GlobalVoting (int leaf_idx, const std::vector< LightSplitInfo > &splits, std::vector< int > *out)
	Perform global voting.
void	CopyLocalHistogram (const std::vector< int > &smaller_top_features, const std::vector< int > &larger_top_features)
	Copy local histgram to buffer.

Detailed Description

template<typename TREELEARNER_T>
class LightGBM::VotingParallelTreeLearner< TREELEARNER_T >

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.

Member Function Documentation

CopyLocalHistogram()

template<typename TREELEARNER_T >

void LightGBM::VotingParallelTreeLearner< TREELEARNER_T >::CopyLocalHistogram ( const std::vector< int > &  smaller_top_features, const std::vector< int > &  larger_top_features  )

protected

Copy local histgram to buffer.

Parameters

smaller_top_features	Selected features for smaller leaf
larger_top_features	Selected features for larger leaf

GlobalVoting()

template<typename TREELEARNER_T >

void LightGBM::VotingParallelTreeLearner< TREELEARNER_T >::GlobalVoting ( int  leaf_idx, const std::vector< LightSplitInfo > &  splits, std::vector< int > *  out  )

protected

Perform global voting.

Parameters

leaf_idx	index of leaf
splits	All splits from local voting
out	Result of gobal voting, only store feature indices

---

The documentation for this class was generated from the following files:

• External/LightGBM_2.2.3/LightGBM-2.2.3/src/treelearner/parallel_tree_learner.h
• External/LightGBM_2.2.3/LightGBM-2.2.3/src/treelearner/voting_parallel_tree_learner.cpp