hidden-markov-app

We implemented this app for our final project in the Stochastic Processes course taught by Prof. Amr Amin.
We had an assignment to implement Baum–Welch Algorithm. The python implementation can be found in hmm_implementation.py file. We then created a PyQT5 app to leverage this implementation.

Application Overview

The main window of our application has two tabs.

The first is a small Help tab, with a simple presentation video that includes an overview of Hidden Markov Models.

The second is the Model tab, where you can use the Hidden Markov Model. There are 5 possible operations that you can choose from the dropdown: Forward Encoding, Backward Encoding, Viterbi Decoding, Baum-Welch Learning, and drawing a transition graph.

Here's how we ran these operations on this sample data.

Number of possible hidden states: 2
Number of possible observations: 2

This was the Transition Graph, and the Forward, Backward, and Viterbi output.

Then, after running the Baum-Welch algorithm, and inputting the updated data into the application, we can recalculate to get the following output.

How to Use

To run this app, you need to have Git, Python, and Graphviz installed on your computer.

Here's how you can check:

# Check Git installation
$ git --version

# Check Python installation
$ python --version

# Check Graphviz installation
$ dot -V

If you have everything installed, you can go ahead and run the following:

# Clone the repository
$ git clone https://github.com/verinak/hidden-markov-app

# Navigate into cloned directory
$ cd hidden-markov-app

# Install app requirements
$ pip install -r requirements.txt

# Run app
$ python run_app.py

If everything's working fine, you should see the app's welcome window.

Contributors

Ola Mamdouh
Verina Michel
Marly Magdy
Maria Anwar
Mirna Tarek
Mariem Nasr