In "Data Modelling", we look at a range of statistical models for data. A "Linear regression" model for (x,y) scatterplot data is an example of a statistical model that you may have met. It has two unknown model parameters that we need to estimate: the straight line’s intercept and slope.
Another example of a model-based technique is the unpaired t-test, where the underlying means of two groups are the unknown parameters of interest. In "Data Modelling" we meet two ways of learning about or estimating unknown model parameters from data. The first is referred to as the frequentist framework and is the one that is usually used in statistical modelling. It leads to hypothesis tests and confidence intervals. The second way is the Bayesian framework which allows us to use existing knowledge about model parameters available from previous studies or experiments. The Bayesian framework requires a lot of computation. In this module you will meet computational techniques that are important to other areas of mathematics and science, and we will use state-of-the-art software "Stan".
Recent courseworks in this module have involved fitting models to Covid-19 cases, comparing engine performance under different environmental conditions, and using historic data to assess how life expectancy has increased over time.