plusml::BinarySVM Class Reference

Basic binary classification SVM. More...

#include <binary_svm.h>

Public Member Functions
	BinarySVM (uint64_t features, bool bias_enabled=true)
	Constructor for BinarySVM class.
uint64_t	Features () const
	Get number of features in each sample for the model.
Eigen::MatrixXf	Parameters () const
	Get current parameters of the model.
void	SetParameters (const Eigen::MatrixXf &parameters)
	Set desired parameters for the model.
void	FitSGD (const Eigen::MatrixXf &samples, const Eigen::MatrixXf &targets, float learning_rate, float l2_alpha, uint64_t batch_size, uint64_t epochs)
	Fit model using stochastic gradient descent with given hyperparameters.
Eigen::MatrixXf	Predict (Eigen::MatrixXf samples, bool sign=true)
	Get predictions for a given matrix of samples.

Detailed Description

Basic binary classification SVM.

Constructor & Destructor Documentation

BinarySVM()

plusml::BinarySVM::BinarySVM ( uint64_t features, bool bias_enabled = true )

explicit

Constructor for BinarySVM class.

Parameters

features	Number of features in each sample
bias_enabled	Specifies whether to use bias or not

Initializes model's parameters to zero

Member Function Documentation

Features()

uint64_t plusml::BinarySVM::Features

(

)

const

Get number of features in each sample for the model.

Returns

Number of features in each sample for the model

FitSGD()

void plusml::BinarySVM::FitSGD	(	const Eigen::MatrixXf &	samples,
		const Eigen::MatrixXf &	targets,
		float	learning_rate,
		float	l2_alpha,
		uint64_t	batch_size,
		uint64_t	epochs )

Fit model using stochastic gradient descent with given hyperparameters.

Parameters

samples	Matrix of samples (MxN where M is number of samples and N is number of features in each sample)
targets	Matrix of targets (Mx1 where M is number of samples)
learning_rate	Learning rate for SGD
l2_alpha	Coefficient for L2 regularization
batch_size	Batch size for SGD
epochs	Number of epochs for SGD

Parameters()

Eigen::MatrixXf plusml::BinarySVM::Parameters

(

)

const

Get current parameters of the model.

Returns

Parameters matrix

Predict()

Eigen::MatrixXf plusml::BinarySVM::Predict	(	Eigen::MatrixXf	samples,
		bool	sign = true )

Get predictions for a given matrix of samples.

Parameters

samples	Matrix of samples (MxN where M is number of samples and N is number of features in each sample)
sign	Apply the Sign function to predictions (default: true)

Returns

Matrix of predicted classes (Mx1 where M is number of samples)

SetParameters()

void plusml::BinarySVM::SetParameters

(

const Eigen::MatrixXf &

parameters

)

Set desired parameters for the model.

Parameters

parameters

Matrix representing parameters to set

---

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

• include/PlusML/binary_svm.h
• src/binary_svm.cpp