A Sequential Monte Carlo (SMC, or "particle filter") Python package for online model estimation and probabilistic forecasting.


What pypfilt can do for you

If there is a system or process that can be:

  • Described (modelled) with mathematical equations; and

  • Measured repeatedly in some (noisy) way.

Then you can use pypfilt to estimate the state and/or parameters of this system.

The Getting Started guide introduces the Lorenz-63 system of ordinary differential equations, which has chaotic solutions for some parameter values and initial states. This guide demonstrates:

  • How to build a simulation model for the Lorenz-63 system;

  • How to generate simulated observations from this model;

  • How to fit the simulation model to observations; and

  • How to generate forecasts that predict the future behaviour of the Lorenz-63 system.

The How-to Guides cover a range of further topics, including how to generate forecasts with:

  • Continuous-time Markov chain (CTMC) models;

  • Discrete-time Markov chain (DTMC) models;

  • Ordinary differential equation (ODE) models; and

  • Stochastic differential equation (SDE) models.

Participating organisations

University of Melbourne



Rob Moss