This is a collection of software modules and scripts that help solve various Machine Learning problems. I will try to focus on problems that are difficult to find in standard library packages.
I will add theory and code for the following problems:
- Linear Regression using l1 (lasso) penalty - A hard problem because of its obvious non-differentiability.
- Sparse Covariance Estimation for a multivariate Gaussian - Estimation of the covariance matrix of a Gaussian using samples drawn from the (unknown) Gaussian. A seemingly standard problem that is surprisingly hard.
- Logistic Regression using l1 (lasso) penalty - self-explanatory.
- Implementation of the Baum-Welch algorithm for Hidden Markov Models (HMMs) - Learning HMM parameters when no tagged data is available.
If you would like me to write up MATLAB or Python code for a Machine Learning algorithm, please don't hesitate to reach out at [email protected] or [email protected]