mbeps/machine-learning-labs-questions

Machine Learning Labs for CS3920 (Royal Holloway University)

CS3920 Machine Learning Labs

Welcome to the repository for the CS3920 Machine Learning Labs. This series of labs is designed to provide a deep dive into the world of machine learning, exploring various algorithms, preprocessing techniques, optimization methods, and performance evaluation strategies.

Each lab corresponds to a week of study and practice, covering specific topics that build upon each other to enhance your understanding of machine learning concepts and applications.

Weekly Lab Topics

Week 1: Further Introductions and Nearest Neighbours

• In-depth introduction to the machine learning course.
• Fundamental machine learning algorithm - Nearest Neighbours.
• Introduction to Conformal Predictors.

Week 2: Introduction to Conformal Prediction

• Study of reliable prediction algorithms with guaranteed reliability.
• Delving into conformal prediction techniques.

Week 3: Conformal Prediction

• Detailed exploration of full conformal prediction.
• Application and analysis of conformal prediction algorithms.

Week 4: General Principles

• Understanding overfitting, underfitting, and the learning curve.
• Exploring the general principles of machine learning.

Week 5: Linear Regression

• Discussion of Least Squares.
• Improvement to Ridge Regression and introduction to Lasso.

Week 6: A Few Advanced Topics

• Data preprocessing and its impact on prediction quality.
• Parameter selection methods.
• Introduction to inductive conformal prediction.

Week 7: Kernels

• Enhancing linear methods with Kernels.
• Application of kernels to various machine learning methods.

Week 8: Two Powerful Algorithms

• Exploration of neural networks.
• Introduction to support vector machines.

Week 9: Two More Advanced Topics

• Understanding and creating pipelines.
• Efficient version of conformal prediction: Cross-conformal predictors.

Week 10: Finale

• Overview of a variety of new prediction algorithms.
• Broad introduction to advanced machine learning techniques.

Running Notebook Locally

These are simple steps to run the notebook locally. Jupyter is required.

1. Clone the Project Locally

git clone [email protected]:mbeps/Machine-Learning-Labs-Questions.git

2. Set Up Environment

Using Anaconda (Preferred)

If you have Anaconda installed, create a new environment and activate it. Once active, install the required packages.

Using Poetry (Alternative for non-Anaconda users)

1. Ensure you have Python 3.10 installed.
2. If you are using Poetry and do not have Anaconda, install the dependencies:

poetry install

3. Run the Project

After setting up the environment and installing all dependencies, navigate to the project's root directory and run the main notebook.

Note: Adjust the specific steps, commands, or any other requirements based on the nature of your project or any additional configurations that might be needed.