Data Science

AIMS

Machine Learning, a core discipline in data science, is prevalent across Science, Technology, the Social Sciences, and Medicine; it drives many of the products we use daily such as banner ad selection, email spam filtering, and social media newsfeeds. Machine Learning is concerned with making accurate, computationally efficient, interpretable and robust inferences from data. Originally borne out of Artificial Intelligence, Machine Learning has historically been the first to explore more complex prediction models and to emphasise computation, while in the past two decades Machine Learning has grown closer to Statistics gaining firm theoretical footing.

This subject aims to introduce undergraduate students to the intellectual foundations of machine learning, and to introduce practical skills in data analysis that can be applied in graduates' professional careers.

CONTENT

Topics will be selected from: prediction approaches for classification/regression such as k-nearest neighbour, naïve Bayes, discriminative linear models, decision trees, Support Vector Machines, Neural Networks; clustering methods such as k-means, hierarchical clustering; probabilistic approaches; exposure to large-scale learning.

Linear models are central to the theory and practice of modern statistics. They are used to model a response as a linear combination of explanatory variables and are the most widely used statistical models in practice. Starting with examples from a range of application areas this subject develops an elegant unified theory that includes the estimation of model parameters, quadratic forms, hypothesis testing using analysis of variance, model selection, diagnostics on model assumptions, and prediction. Both full rank models and models that are not of full rank are considered. The theory is illustrated using common models and experimental designs.

Modern applied statistics combines the power of modern computing and theoretical statistics. This subject considers the computational techniques required for the practical implementation of statistical theory, and includes Bayes and Monte-Carlo methods. The subject focuses on the application of these techniques to generalised linear models, which are commonly used in the analysis of categorical data.

AIMS

This capstone subject for the Data Science major combines statistical reasoning and practical computing skills to solve challenging problems with big data.

INDICATIVE CONTENT

Students will learn about communication of quantitative information and insights; presentation skills; report writing; project management; problem formulation using case studies; data collection and measurement protocols; data from surveys and experiments; issues in capturing and dealing with “big data”; dimension reduction; data visualisation; fitting formulated models to data to infer insightful information about populations; ethics in quantitative research; working effectively in teams.