A well-designed experiment is an efficient way of learning about the world. Typically, an experiment may involve varying several factors and observing the value of a response at settings of combinations of values of these factors. The mathematical challenge is then to choose which settings to use in order to gain the maximum information from the resulting data.
Experiments are performed in all branches of science, engineering and industry. In recent years, traditional application areas such as agriculture, manufacturing, medicine and pharmaceutical science
have been joined by bioinformatics, genetics, drug discovery, finance and economics. Problems of increasing size and complexity from these new areas have led to the development of many new
methods for designing and analysing experiments. The aim of this module is to provide a grounding in the statistical and mathematical methods that underpin the design and analysis of experiments, before exploring a number of areas where recent and ongoing developments are taking place. Mathematical criteria for quantifying the information available from a given design will be defined and explored, and will underpin much of the material in the module.
Examples from a range of application areas will be used to motivate and illustrate the methods. The R statistical computing language will be used for the practical design and analysis of exemplar experiments.
Although MATH2010 and MATH2011 are listed as pre-requisites, interested students who have not taken these modules are encouraged to discuss the module with the Current Module Co-ordinator.
Pre-requisites: MATH2010 AND MATH2011