knngo

Go library to machine learning classification problems

Getting Started

Run at terminalgo get -u github.com/ai-brasil/knngo

Import the packageimport "github.com/italojs/knngo"

Prerequisites

1º: Only the class column can be a NaN (Not a Number)

2º: The column class must be the last column.

3º: When you read your file (e.g. your .CSV file) don't worry to transform the type of your data to number, you can use the values as string.

Methods

knn.PrepareDataset(percent float32, records [][]string) (train [][]string, test [][]string, err error)

func Classify(trainData [][]string, dataToPredict []string, k int) (result string, err error)

Example

package main

import (
	"fmt"
	"github.com/italojs/knngo"
)

func main (){
	records := readFile("../datas/wdbc.csv")
	
	// Use the PrepareDataset method is optional
	percentToTrain := float32(0.6)
	train, test, err := knn.PrepareDataset(percentToTrain, records)
	if err != nil{
		log.Fatalln(err)
	}

	// You can choose some not classificated data to classify
	// The line 5 of test dataset was choosed randomicly by programmer just to example
	result, err := knn.Classify(train, test[5], 10)
	if err != nil{
		log.Fatalln(err)
	}

	fmt.Println(result)
}

Authors

*Italo Jose - Initial work - italojs

License

This project is licensed under the MIT License - see the LICENSE.md file for details