SUFTware

Statistics Using Field Theory

Example density estimate using WHO data.

SUFTware is a lightweight Python package that provides provides fast and robust implementations of Bayesian Field Theory (BFT) methods for low-dimensional statistical inference. BFT is a grid-based approach to Bayesian nonparametric inference. By using a grid in lieu of specific stochastic processes (such as Dirichlet processes or Gaussian processes), BFT allows certain types of problems to be solved in a fully Bayesian manner without requiring any large-data approximations.

Currently, SUFTware supports a one-dimensional density estimation called DEFT. DEFT has substantial advantages over standard density estimation methods, including, including kernel density estimation and Dirichlet process mixture modeling. See [Chen et al., 2018; Kinney 2015; Kinney 2014].

Installation

pip install suftware

Requirements

  • Python >= 3.6.3
  • numpy >= 1.13.3
  • scipy >= 1.0.0
  • matplotlib >= 2.1.0

Quick Start

import numpy as np
import suftware as sw

# Generate random data
data = np.random.randn(100)

# Perform one-dimensional density estimation using SUFTware
density = sw.Density(data)

# Visualize results
density.plot()

Indices and tables