Reinforcement-Learning-CSL348

About the Repository

This repository hosts several projects showcasing various reinforcement learning algorithms and statistical analysis techniques. These projects were developed as part of my course at Northcap University, where I learned about these topics in-depth.

Key Projects and Algorithms

1. BlackJack with First-Visit MC:

   • Algorithm: First-Visit Monte Carlo
   • Description: Implements the First-Visit Monte Carlo method for learning the value function in a Blackjack environment.

2. Calculating Probabilities:

   • Algorithm: Probability Calculations
   • Description: Focuses on calculating probabilities, fundamental in reinforcement learning.

3. Correlation Analysis Projects:

   • Algorithm: Correlation Analysis
   • Description: Analyzes relationships between different variables in datasets.

4. Q-Learning Implementation:

   • Algorithm: Q-Learning
   • Description: Implements the Q-Learning algorithm for model-free reinforcement learning.

5. Thompson Sampling:

   • Algorithm: Thompson Sampling
   • Description: Implements the Thompson Sampling algorithm for multi-armed bandit problems.

6. Upper Confidence Bound (UCB):

   • Algorithm: UCB
   • Description: Implements the UCB algorithm for balancing exploration and exploitation in multi-armed bandit problems.

7. Markov Chains:

   • Algorithm: Markov Chains
   • Description: Explores Markov Chains, a key concept in reinforcement learning.

8. Policy and Value Iteration:

   • Algorithms: Policy Iteration and Value Iteration
   • Description: Implements Policy Iteration and Value Iteration algorithms for finding the optimal policy.

Datasets Used

1. Ads Optimisation Dataset (Ads_Optimisation.csv):

   • Used in the Thompson Sampling and UCB projects.
   • Contains data used to simulate ad selection scenarios for optimizing ad displays.

2. Student Performance Dataset (Math) (student-mat.csv):

   • Used in correlation analysis projects to study relationships between different attributes and student performance in math courses.

3. Student Performance Dataset (Portuguese) (student-por.csv):

   • Similar to the math dataset, used to study relationships between different attributes and student performance in Portuguese courses.

Education

All the concepts and algorithms implemented in these projects were taught as part of my course at Northcap University. The course provided a comprehensive understanding of reinforcement learning and statistical analysis, which I have applied in these projects.