PGM - Bayesian Networks and Markov Models for handwiriting samples, specifically for 2 datasets - ‘th’ and ‘and’ dataset.
The purpose of this project is to document the output and analysis of Determining Probabilities of Handwriting Formations using PGMs for ‘th’ and ‘and’ dataset. Task 1 consists of finding which features are independent of which features for ‘th’ dataset. Task 2 involves creating different Bayesian networks and finding the one with the best K2 score for ‘th’ dataset. Task 4 consists of the same task as Task 2, but with ‘and’ dataset. Finally, Task 3 consists of converting the above found, best Bayesian networks to Markov Models and comparing their results.\
- The code is present in the "main.py" file
- The "Report.pdf" file details the project and the outcome.\
- The excel files contain the CPDs.