JKLearn (Machine Learning)

Data Loading (CSV)

JKLearn provides a streamlined way to load datasets directly into the environment using the C-optimized basic module.

spec_manifest.jk

import basic

let data = Csv("data.csv").load()
println(data)

Naive Bayes Classification

A probabilistic classifier based on Gaussian distribution, designed for fast inference on structured data.

spec_manifest.jk

import supervised
import basic

let data = Csv("a.csv").load()
let labels = [1, 1, 0, 0]
let dataset = Data(data, labels)

let nb = NaiveBayes()
    .fit(dataset)
    .predict([[10, 100]])

println(nb)

K-Nearest Neighbors (KNN)

Distance-based classification using optimized Euclidean distance computation, suitable for academic research and SSE tracking.

spec_manifest.jk

import basic
import supervised

let data = [[10, 20], [20, 30], [30, 40], [40, 50]]
let labels = [1, 1, 0, 0]
let dataset = DataSet(data, labels)

let knn = Knn(3)
    .fit(dataset)
    .predict([[10, 21]])

println(knn)

High-Performance Tensors

Native C-core tensor operations supporting reshaping and vectorized mathematical transformations without external dependencies.

spec_manifest.jk

import Tensor

let raw = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0]

let t = Tensor<Float>(raw, [10])
    .reshape([2, 5])
    .add(5.0)
    .mul(2.0)
    .exports()

println(t)

← Return to Index Technical Documentation