Welcome to Jajapy’s documentation!

jajapy is a python library implementing the Baum-Welch algorithm on various kinds of Markov models.

Note

This project is under active development.

Installation

To use jajapy, first install it using pip:

pip install jajapy

Note

jajapy runs on Windows, Linux and MacOS. However, it is not multithreaded on Windows and MacOS: hence, we trully recommend to use it on Linux!

Selected Features

• Learning HMMs, MCs, MDPs, CTMCs and GoHMMs from traces.

• Parameter estimation for synchronous composition of CTMCs.

• Compatibility with Prism and Storm.

Content

• References
  • Models
  • Learning Algorithms
  • Others
• Library description
  • 1. Installation
  • 2. Workflow
  • 3. Models
  • 4. Jajapy and Stormpy
  • 5. Jajapy and Prism
  • 6. Examples and tutorial
• Tutorial/Examples
  • 1. A simple example with MC: Hello World
  • 2. Learning an MC with random restart
  • 3. Learning an MDP from a prism file
  • 4. Learning an MDP using Active-BW
  • 5. Learning CTMCs
  • 6. Parameter estimation for PCTMCs
  • 7. A simple example with HMM
  • 8. Learning a GoHMM from a csv
  • 9. Learning an MC with Alergia
  • 10. Learning an MDP with IOAlergia
• Formalism

Source Code

The source code is available on github.

Contact

raphal20 at ru dot is