📦 ml_knn

A production-ready K-Nearest Neighbors (KNN) classifier implemented in pure Dart.
Supports Euclidean, Manhattan, and Cosine distances, with optional weighting and normalization.
Perfect for lightweight ML tasks on mobile, web, or server-side Dart.

✨ Features

• 🧠 Distance metrics: Euclidean, Manhattan, Cosine
• ⚖️ Optional weighted voting
• 📊 Built-in normalization support
• 🔁 Batch predictions
• 💾 Model serialization/deserialization
• 🧪 Full test coverage

🚀 Quick Start

import 'package:ml_knn/ml_knn.dart';

void main() {
  final features = [
    [1.0, 0.0],
    [0.0, 1.0],
    [1.0, 1.0],
  ];
  final labels = ['A', 'B', 'A'];

  final knn = KNN(k: 3, normalize: true);
  knn.fit(features, labels);

  final prediction = knn.predict([1.2, 0.1]);
  print('Prediction: $prediction');

  final probs = knn.classProbabilities([1.2, 0.1]);
  print('Probabilities: $probs');
}