XGBoostPredictor Class Reference

Simple wrapper around XGBoost C API. More...

#include <XGBoostPredictor.h>

List of all members.

Public Member Functions
	XGBoostPredictor (std::string const &path, int numFeatures)
	Constructor.
float	Predict (float const *x) const
	Computes prediction of the XGBoost model for a single example.
Static Private Member Functions
static void	CheckCall (std::string const &functionName, int exitCode)
	Throws XGBoostPredictorException if the exit code is not 0.
Private Attributes
BoosterHandle	booster_
	Pointer to XGBoost model.
int	numFeatures_
	Number of input features specified in the constructor.

---

Detailed Description

Simple wrapper around XGBoost C API.

This class loads a saved XGBoost model from a file and computes its predictions, for one example at a time. The predictions require memory allocation and deallocation for each example, which is not optimal. Unfortunately, in XGBoost 0.90 API this is the only way to compute predictions per example.

---

Constructor & Destructor Documentation

XGBoostPredictor::XGBoostPredictor	(	std::string const &	path,
		int	numFeatures
	)

Constructor.

Parameters:

[in]	path	Path to a file with XGBoost model.
[in]	numFeatures	Number of input features.

---

Member Function Documentation

static void XGBoostPredictor::CheckCall	(	std::string const &	functionName,
		int	exitCode
	)			[static, private]

Throws XGBoostPredictorException if the exit code is not 0.

Arguments are forwarded to XGBoostPredictorException.

float XGBoostPredictor::Predict

(

float const *

x

)

const

Computes prediction of the XGBoost model for a single example.

Parameters:

[in]

x

Array of features describing the example. Starting from the given position, numFeatures_ elements will be read.

Returns:

Prediction of the model.

---

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

• include/XGBoostPredictor.h
• src/XGBoostPredictor.cc