pyhuge Documentation
pyhuge is a native Python package for high-dimensional undirected graph
estimation and inference workflows inspired by huge.
If you are new, follow this order:
Package capabilities
- Estimators:
huge,huge_mb,huge_glasso,huge_ct,huge_tiger - Selection and preprocessing:
huge_select,huge_npn - Simulation and inference:
huge_generator,huge_roc,huge_inference - Plot helpers:
huge_plot_sparsity,huge_plot_roc,huge_plot_graph_matrix,huge_plot_network,huge_plot - Dataset helper:
huge_stockdata - Environment probe:
pyhuge.test(require_runtime=False),pyhuge-doctor
Runtime model
pyhuge 0.3 is native Python:
- Core logic in Python + NumPy/SciPy
- Native C++ kernels (
pyhuge._native_core) formb/glasso/tiger - No R runtime dependency
Documentation map
- Concept pages: Quick Start, FAQ
- Practical pages: Quick Start, Tutorials, Troubleshooting
- Reference pages: API Reference, Function Manual, Design, Performance