Skip to content

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:

  1. Quick Start
  2. Installation
  3. Beginner Workflow
  4. Troubleshooting

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) for mb/glasso/tiger
  • No R runtime dependency

Documentation map