Welcome to the documentation of pygom

pygom is a package that aims to facilitate the application of ordinary differential equations (ODEs) in the real world, with a focus in epidemiology. This package helps the end user define their ODE system in an intuitive manner and provides convenience functions - making use of various algebraic and numerical libraries in the backend - that can be used in a straight forward fashion.

This is an open source project hosted on Github.

User Documentation

• Getting started
  • What this package does
  • Obtaining the package
  • Testing the package
• Transition Object
  • Defining the equations
  • Model Addition
  • Transition type
• Stochastic representation of ode
  • Stochastic Parameter
  • Continuous Markov Representation
  • Repeatable Simulation
• Convert ODE into transitions
  • Simple Problem
  • ODE With Birth and Death Process
  • Hard Problem
• Solving Boundary Value Problems
  • Simple model 1
  • Simple model 2
• Example: Parameter Estimation 1
  • Estimation under square loss
    • SIR Model
    • Difference in gradient
    • Optimized result
• Pre-defined Example common_models

Code Documentation and FAQ

• Frequent asked questions
  • Code runs slowly
  • Why not compile the numeric computation form sympy against Theano
  • Why not use mpmath library throughout?
  • Computing the gradient using SquareLoss is slow
  • Can you not convert a non-autonumous system to an autonomous system for me automatically
  • Getting the sensitivities from SquareLoss did not get a speed up when I used a restricted set of parameters
  • Why do not have the option to obtain gradient via complex differencing
• Code documentations
  • model
    • common_models
    • transition
    • deterministic
    • stochastic
    • epi_analysis
    • ode_utils
  • loss
    • ode_loss
    • calculations
    • confidence_interval
    • loss_type
    • get_init

References

• References

Indices and tables

• Index
• Module Index