Durham University
Programme and Module Handbook

Postgraduate Programme and Module Handbook 2021-2022 (archived)

Module MATH52115: Advanced Statistics and Machine Learning: Regression and Classification

Department: Mathematical Sciences

MATH52115: Advanced Statistics and Machine Learning: Regression and Classification

Type Tied Level 5 Credits 15 Availability Available in 2021/22 Module Cap None.

Prerequisites

  • Advanced Statistical and Machine Learning: Foundations and Unsupervised Learning

Corequisites

  • None

Excluded Combination of Modules

  • None

Aims

  • Provide advanced knowledge and critical understanding of the paradigms and fundamental ideas of Bayesian statistics and machine learning.
  • Provide advanced knowledge and critical understanding of the methodology and applications of Bayesian statistics and machine learning.

Content

  • Advanced statistical modelling (e.g. linear models, graphical models, Gaussian processes).
  • Advanced supervised machine learning (e.g. random forests, boosting, deep neural networks, deep statistical models).

Learning Outcomes

Subject-specific Knowledge:
  • Advanced understanding of statistical modelling frameworks and methods.
  • Advanced understanding of supervised machine learning frameworks and methods.
Subject-specific Skills:
  • Ability to use Bayesian theory and inference to frame, analyse, and formalize practical problems, and to reflect critically upon this use.
  • Ability to select and apply appropriate computationally-intensive methods to practical problems, and to reflect critically upon their application.
  • Ability to select or to develop, and to apply, appropriate models to practical problems, and to reflect critically upon their application.
  • Ability to select, adapt, and apply appropriate machine learning methods to practical problems, and to reflect critically upon their application.
Key Skills:

    Modes of Teaching, Learning and Assessment and how these contribute to the learning outcomes of the module

    • Lectures demonstrate what is required to be learned and the application of the theory to concrete examples.
    • Practical classes concretize understanding via the application of calculational and computational methods to more complex problems, as well as providing feedback and encouraging active engagement.
    • Coursework will assess students' ability to implement calculational and computational methods on both synthetic and real problems.

    Teaching Methods and Learning Hours

    Activity Number Frequency Duration Total/Hours
    Lectures on Regression 12 3 per week, weeks 16 - 19, term 2 1 hour 12
    Practical classes on Regression 4 1 per week, weeks 16 - 19, term 2 1 hour 4
    Lectures on Classification 12 3 per week, weeks 16 - 19, term 2 1 hour 12
    Practical classes on Classification 4 1 per week, weeks 16 - 19, term 1 hour 4
    Preparation, reading, and self-study 118

    Summative Assessment

    Component: Coursework Component Weighting: 100%
    Element Length / duration Element Weighting Resit Opportunity
    Coursework on Regression 5 weeks 50%
    Coursework on Classification 5 weeks 50%

    Formative Assessment:


    Attendance at all activities marked with this symbol will be monitored. Students who fail to attend these activities, or to complete the summative or formative assessment specified above, will be subject to the procedures defined in the University's General Regulation V, and may be required to leave the University