The repo contains Assignments and a Project, as a part of the course Machine Data and Learning 2021.
- A1: Bias and Variance
Training mulitiple models with different polynomial dgrees, and finding the optimal degree by comparing the bias and variance for each case. - A2: Value Iteration and Markov Decision Process (MDP)
- Part 1: Manual calculation for Value Iteration, using Bellman Update Equation.
- Part 2: Finding optimal MDP policy by using Value Iteration algorithm and Bellman Update Equation.
- Part 3: Linear Programming.
Finding an optimal MDP policy by usingcvxpy
Linear Programming Solver python library.
- A3: Belief States and Partially Observable Markov Decision Process (POMDP)
- Part 1: Manual steps for Belief State calculation.
- Part 2: Finding optimal POMDP policy by using Sarsop solver.
An implementation of Genetic Algorithm for manipulating sets of 11-dimensional vectors to optimally fit the training data.