N-Arm Spiral Dataset

N-Arm Spiral Dataset Overview

Dataset Purpose

• Generate a toy dataset that cannot be classified by a linear classifier.
• Extend from the 2‑Arm Spiral dataset to support more than two arms.

Dataset Applications

• Demonstrate the difference between linear and nonlinear classifiers.
• Provide simple shallow multilayer perceptron (MLP) classification examples, comparing linear functions and nonlinear activation functions (e.g., ReLU) in hidden layers.

Dataset Construction Tools

• Use numpy to construct the dataset.
• Use matplotlib for visualization.
• Use PyTorch for classifier implementation.

License

• This project follows the BSD license.