The objective is to provide the theoretical foundations for quantitative time series analysis and machine learning. Participants will receive an introduction to quantitative techniques for visualising, analysing, modelling, classifying and forecasting time series. The course will provide a multidisciplinary approach based on methodologies and techniques from the fields of engineering, mathematics, statistics, physics and econometrics. It will demonstrate how to model the dynamics of complex systems using empirical observations, test for significant patterns and identify predictive signals to formulate automated decision support tools. Starting from the basics of describing time series as being generated by a stochastic process, the course will present the shortfalls of traditional techniques when dealing with real-world data. The challenges of modelling data with non-normal distributions, nonlinear relationships, and regime-switching behaviour and structural breaks will be explored. Methods for managing different types of uncertainty in the modelling process will be introduced and the importance of providing multiple scenarios for decision-making will be investigated.
Pre-requisites: Mathematics, statistics, probability, programming