bayesian

README

bayesian

A simple, efficient binary bayesian classification library

Documentation

Overview

Package bayesian implements a bayesian classifier

Index

• Constants
• Variables
• type BinaryClassifier
  • func NewBinaryClassifier(smoothingFactor float64) (BinaryClassifier, error)
• type Classifier
  • func NewClassifier(categories int, smoothingFactor float64) (Classifier, error)

Constants

const Negative = 0

Negative represents the negative category in a binary classifier

const Positive = 1

Positive represents the positive category in a binary classifier

Variables

var ErrInvalidSmoothingFactor = errors.New("bayesian: invalid smoothing factor")

ErrInvalidSmoothingFactor is an error we throw when the smoothing factor provided is less than 0

Types

type BinaryClassifier

type BinaryClassifier interface {
	Scores(doc []string) ([]*big.Float, int, bool)
	LearnPositive(doc []string) error
	LearnNegative(doc []string) error
}

BinaryClassifier is an interface to a simple bayesian binary classifier

func NewBinaryClassifier

func NewBinaryClassifier(smoothingFactor float64) (BinaryClassifier, error)

NewBinaryClassifier creates a new bayesian classifier with two classes

type Classifier

type Classifier interface {
	Scores(doc []string) ([]*big.Float, int, bool)
	Learn(doc []string, category int) error
}

Classifier is an interface to a general multi-class bayesian classifier

func NewClassifier

func NewClassifier(categories int, smoothingFactor float64) (Classifier, error)

NewClassifier creates a new instance of a bayesian with n classes classifier