knn

Documentation

Index

• func EuclideanDistance(p1, p2 []float64) float64
• type Classifier
  • func Load(directory string) (classifier *Classifier, err error)
  • func NewClassifier(wordVecConfig wordvectors.Config, params map[string]interface{}) *Classifier
  • func NewClassifierFromFile(name string) (*Classifier, error)
  • func (c *Classifier) Learn(texts dataset.DataSet, pipe *pipeline.Config) float32
  • func (c *Classifier) Predict(text string, pipe *pipeline.Config) (predictedClass string, proba float32)
  • func (c *Classifier) Save(directory string) error
  • func (c *Classifier) SaveToFile(name string) error
• type KNN
  • func NewKNNClassifierFromFile(name string) (*KNN, error)
  • func (knn *KNN) PredictMany(x [][]float64) (predictedLabels []string, probabilities []float64)
  • func (knn *KNN) PredictOne(x []float64) (predictedLabel string, probability float64)
  • func (knn *KNN) SaveToFile(name string) error
• type Parameters

Functions

func EuclideanDistance

func EuclideanDistance(p1, p2 []float64) float64

EuclideanDistance calculates euclidean distance between two points

Types

type Classifier

type Classifier struct {
	KNN       *KNN
	VectorMap *wordvectors.VectorMap
	K         int
}

Classifier is a K-Nearest Neighbors classifier

func Load

func Load(directory string) (classifier *Classifier, err error)

func NewClassifier

func NewClassifier(wordVecConfig wordvectors.Config, params map[string]interface{}) *Classifier

NewClassifier creates a KNN classifier with truncate and file data

func NewClassifierFromFile

func NewClassifierFromFile(name string) (*Classifier, error)

func (*Classifier) Learn

func (c *Classifier) Learn(texts dataset.DataSet, pipe *pipeline.Config) float32

Learn takes the training texts and trains the K-Nearest Neighbors classifier

func (*Classifier) Predict

func (c *Classifier) Predict(text string, pipe *pipeline.Config) (predictedClass string, proba float32)

Predict predict a class for a given text

func (*Classifier) Save

func (c *Classifier) Save(directory string) error

Save persists the model to a file

func (*Classifier) SaveToFile

func (c *Classifier) SaveToFile(name string) error

type KNN

type KNN struct {
	K      int
	Data   [][]float64
	Labels []string
}

KNN main structure

func NewKNNClassifierFromFile

func NewKNNClassifierFromFile(name string) (*KNN, error)

func (*KNN) PredictMany

func (knn *KNN) PredictMany(x [][]float64) (predictedLabels []string, probabilities []float64)

PredictMany performs a classification on multiple input vectors

func (*KNN) PredictOne

func (knn *KNN) PredictOne(x []float64) (predictedLabel string, probability float64)

PredictOne performs a classification on one input vector

func (*KNN) SaveToFile

func (knn *KNN) SaveToFile(name string) error

type Parameters

type Parameters struct {
	K int `mapstructure:"k"`
}

Parameters represents the model hyperparameters