It includes my codes for topics as:
- Fitting k-means clustering by minimizing variance/objective function
- General Stats Optimization topics such as maximization of different likelihood functions: non-linear, constrained, block relaxtion ones, etc
- Expectation Maximization Algorithm for a mixture of Multivariate Gaussians
- Bootstrap and Permutation for linear regression and other problems
- Monte Carlo methods and Importance Sampling
- Markov-Chain Monte Carlo for bayesian stats: Metropolis-Hastings and Gibbs-Sampler for bayesian regression
- Hidden Markov Chains applied to finance
- Some linear algebra topics
- Course Final Work: I've used a Sequential Quadratic Programming (SQP) method for optimizing mixture copula parameters instead of classical EM algos. To assess performance I've conducted a Monte Carlo expectation and s.d. estimation of a simulated mixture. Used the proposed optimization method to a financial portfolio optimization problem.