Statistical Machine learning projects
Party affiliation project:
In this project, you will implement the parameter learning for NB classifiers and use it to compare with logistic regression with LASSO. You will apply these classifiers to predict the party affiliation of either Democrat or Republican of US Congresspeople (the class variable) based on their votes for 16 different measures. Not all congresspeople voted on all 16 measures, so sometimes entries in this dataset will have missing attributes; however, we will still be able to utilize our Bayes Network to accurately classify these examples. To keep things simple, the class and attribute variables are all binary with 0, 1 corresponding to a no and yes vote respectively.
Denoising image:
In an image restoration problem, you are given an image corrupted by noise X and you want to recover the original image Y. You can apply Gibbs sampling to a simple Ising Markov random field model to solve this task.
This is a great project. Can we have the Github link?
Here is the Github link.