Applied Machine Learning Projects

This repository contains two machine learning projects developed for the Applied Machine Learning module at the University of Sussex.

The projects explore applications of machine learning in natural language processing and computer vision.

---

Repository Structure

applied-machine-learning-projects
│
├── task1-spam-classification
│   ├── spam_classification.ipynb
│   ├── spam_results.csv
│   └── README.md
│
├── task2-face-keypoint-detection
│   ├── face_keypoint_detection.ipynb
│   ├── face_results.csv
│   └── README.md
│
└── report
    └── AML-report.pdf

Task 1 – Spam Email Classification

Goal:
Build a machine learning model to classify messages as spam or ham (normal).

Techniques used:

• TF-IDF feature extraction
• Sentence-BERT embeddings
• Logistic Regression

More details can be found in:

task1-spam-classification/README.md

---

Task 2 – Face Keypoint Detection

Goal:
Predict facial landmark coordinates from face images using deep learning models.

Techniques used:

• Transfer learning
• ResNet architectures
• Deep learning regression

More details can be found in:

task2-face-keypoint-detection/README.md

---

Requirements

Python 3.8+

Main libraries used:

• numpy
• pandas
• scikit-learn
• matplotlib
• PyTorch
• sentence-transformers

Install dependencies with:

pip install numpy pandas scikit-learn matplotlib torch sentence-transformers

---

Report

The full experimental report is available in:

report/AML-report.pdf