Welcome to anglepy’s documentation!¶
anglepy is a specialized Python package designed for deep distributional regression with directional and circular data. It provides the deep engression-based architecture, kernels, metrics, and customized loss functions needed to effectively build, train, and evaluate circular data neural networks.
anglepy offers a unified framework to handle the unique geometric properties of angular data within modern machine learning workflows. It offers the following:
Prediction with model-intrinsic uncertainty quantification
Extrapolation on the circle
Sufficient dimension reduction
Testing equality of conditional distributions
Getting Started¶
To install the latest stable release, simply run the following command in your terminal:
pip install anglepy
Explore the Documentation¶
Step-by-step instructions to set up anglepy and its dependencies in your environment.
Detailed documentation of the architectures, kernels, loss functions, and metrics.
A Jupyter notebook demonstrating core workflows and package functionality.
Information on how to cite the paper in your work.