Metadata-Version: 2.4
Name: k-reg
Version: 2.0.0
Summary: High-performance non-linear K-Regressor (v19, predict=stream).
Home-page: https://github.com/romain/regression-k
Author: Romain
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Requires-Dist: numpy>=1.19.0
Requires-Dist: scikit-learn>=0.24.0
Requires-Dist: numba>=0.53.0
Dynamic: author
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: home-page
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

# ⚡ K-Reg (KRegressor)

**The Linear Regression Killer.**
Faster than C-optimized linear algebra. More accurate thanks to non-linearity. Stable on small data.

`KRegressor` is a drop-in replacement for Scikit-Learn's `LinearRegression`.

## 🚀 Benchmark

| Dataset Size | Model | Time | R² Score |
|--------------|-------|------|----------|
| **Small (N=50)** | LinearRegression | 1.0ms | 0.87 |
| | **KRegressor** | **0.5ms** | **0.97** |
| **Huge (N=500k)** | LinearRegression | 380ms | 0.75 |
| | **KRegressor** | **80ms** | **0.92** |

## 📦 Installation

```bash
pip install k-reg==2.0.0
```

## ✅ v2.0.0 notes

- Default inference is `predict_mode="stream"` (no `X_aug` allocation)
- v19 adds interaction autopilot (feature selection + (A+B)/(A-B) interactions)
