Course finder

<< return to browsing

Semester 2

Data Analysis and Machine Learning 4 (ELEE10031)

Subject

Electronics and Electronic Engineering

College

SCE

Credits

10

Normal Year Taken

4

Delivery Session Year

2023/2024

Pre-requisites

Please see the "Other Requirements" box

Course Summary

Data-driven solutions using machine learning are becoming increasingly prevalent in society. It is essential that the modern engineer has the tools to analyse and interpret data, and be able to apply machine learning methods where appropriate. They should also have an appreciation of the ethical issues that can arise when making decisions based on these methods.This course aims to provide engineering students with the skills to process and examine different forms of data in Python, and an understanding of how machine learning methods can use this data to solve classification and regression problems. They will learn how to implement these methods in Python using Scikit-learn and PyTorch. The students will also gain an awareness of: when it is appropriate to use a particular method (if any); best practices; the ethical issues that can occur when sourcing data and deploying machine learning in the real world.

Course Description

This course takes a hands-on approach to performing data analysis and using machine learning. It combines traditional teaching with lab sessions using interactive Jupyter notebooks where students will develop and run Python code to analyse data and try out machine learning methods for themselves. Each week's teaching will consist of a lecture to introduce material and a follow-up lab session to put it into practice. The topics for each week are (provisionally):1. Introduction, data modalities, variable types2. Summarising and visualising data3. Preprocessing data, principal component analysis, clustering4. Machine learning and ethics5. Linear models for regression6. Linear models for classification7. Model selection and evaluation8. Classification and regression trees, bagging and boosting9. Neural networks10. Modern pre-trained networks, and how to use them

Assessment Information

Written Exam 0%, Coursework 100%, Practical Exam 0%

Additional Assessment Information

Mini-tests (50%) - There are 3 mini-tests worth 16.67% each. These open-book tests are taken in-person. Each will consist of (i) short answer questions on theory; (ii) some programming exercises. Coursework 1 (25%) - The student will record a short presentation performing a case study on a real-world application of machine learning.Coursework 2 (25%) - The student will perform exploratory data analysis and machine learning on a given dataset and will produce a report on their findings.

view the timetable and further details for this course