[Goal] LinearRegression: Near-Identical Python→C# Translation

[Nucs]

The README screenshot showcases a LinearRegression implementation - Python on the left, C# on the right.

This is what NumSharp was built for: enabling near-identical code translation from Python to C#.

What This Demonstrates

The screenshot proves NumSharp's core promise: you can take NumPy code and translate it to C# with minimal changes:

Python/NumPy	C#/NumSharp
np.hstack(...)	np.hstack(...)
np.mean(X, 0)	np.mean(X, 0)
np.std(X, 0)	np.std(X, 0)
np.expand_dims(y, -1)	np.expand_dims(y, -1)
np.zeros(...)	np.zeros(...)
X @ self.params	np.matmul(X, @params)
(y - y_pred)**2	np.power(y - y_pred, 2)

The entire LinearRegression class - constructor, fit(), score(), and predict() methods - translates almost 1-to-1.

Achievement Unlocked 🏆

This example should be preserved, tested, and maintained as a canonical demonstration of NumSharp's capabilities.

Tasks

• Add LinearRegression example to test suite to ensure it never breaks
• Verify all APIs used still work correctly with NumPy 2.x behavior
• Consider adding to documentation as a "getting started" example