Module Descriptor School of Computer Science and Statistics

2 lecture/lab hours per week

Students who complete this module should be able to:

• DS3LO1 Define a Markov chain and describe its theory
• DS3LO2 Identify an appropriate Monte Carlo simulation method for a given probability distribution and implement it.
• DS3LO3 Describe and implement the state of the art methodology in several topical applications in data science.
• DS3LO4 Complete a data science project that applies the methods of this and other modules to a real data set.

This module continues on from CS7CS4 (Machine Learning) with a focus on sampling methods and topical applications.  It also gives an opportunity for students to apply, through a large project, the methods that they have explored in CS7DS1 (Data Mining & Analytics) and that they are currently exploring in CS7DS2 (Optimisation Algorithms for Data Analysis).

Markov chains;

Monte Carlo sampling methods;

Hierarchical graphical models;

Introduction to databases: MySQL, and tidyr.

Project: application of statistical and machine learning methods to real data example.

Bishop, C.M., “Pattern Recognition and Machine Learning,” Springer-Verlag New York, 2006.

Murphy, Kevin P., “Machine Learning: A Probabilistic Perspective,“ MIT Press, 2013.

Wood, S. “Core Statistics,” Cambridge University Press, 2016.

Coursework: 100%

30% of the coursework mark will be allocated to smaller assignments and 70% to a larger-scale project to be handed in at end of module.

Assessment in the Supplemental session will be based on 100% coursework.