Introduction to Machine Learning
Module Aims: This module aims to introduce students to some foundational ideas in machine learning, while familiarising them with a set of canonical methods and algorithms
Module Learning Outcomes: By the end of the module, students should be able to:
- Describe modelling assumptions, algorithms and analyses using the terminology of machine learning
- Identify and apply appropriate machine-learning methods to solve a range of inferential and prediction problems and critically compare the stability of results obtained using standard approaches
- Recognise contexts in which algorithms for inference and prediction derived from flexible high dimensional models can offer advantages over classical statistical methods
- Appreciate that flexible modelling can only lead to useful out-of-sample prediction when it is possible to impose valid structural assumptions about the data-generating process
- Identify useful objective function penalisations corresponding to a variety of such structural assumptions
- Interpret the output of machine learning algorithms in the context of the underlying modelling assumptions.
Pre-requisites: Advanced Biostatistics for HDS module. Specifically, familiarity with parametric regression models, methods for model comparison/choice, regularisation/penalisation, dimension reduction techniques, latent structures and cluster analysis.
Teaching Strategy: Lectures and computer practicals. Some preliminary reading may be required.
Assessment: Practical analysis with assessed report. The students will receive one dataset, and a set of questions focusing on two standard ML methods. 40% of the module grade will be assigned to each ML method; 20% of the module grade will be assigned to an open-ended discussion.
Module Length and Dates: 4 days