This course considers statistical techniques to evaluate processes occurring through time. It introduces students to time series methods and the applications of these methods to different types of data in various contexts (such as actuarial studies, climatology, economics, finance, geography, meteorology, political science, risk management, and sociology). Time series modelling techniques will be considered with reference to their use in forecasting where suitable. While linear models will be examined in some detail, extensions to non-linear models will also be considered.

Coverage will begin with basic concepts on difference equations, which underlie all of the time-series methods employed in the rest of the course. Some basic rules of matrix algebra will also be reviewed, as they are essential for the study of multiple time series models.  We will then proceed with the traditional Box-Jenkins approach to time series analysis and then continue with dynamic modeling and regression approaches. Non-linear models such as conditionally heteroskedastic (time-varying variance) models will also be studied. In addition, recent developments such as cointegration analysis, error correction models, vector autoregression, and multivariate vector error correction will also be covered. We will learn not only how to construct these models but also how to use them in applied analysis.

Heavy emphasis will be given to fundamental concepts and applied work. Since this is a course on applying time series techniques, different examples will be considered whenever appropriate.

 Upon successful completion of the requirements for this course, students will be able to:

• Understand and apply the concept of stationarity to the analysis of time series data in various contexts (such as actuarial studies, climatology, economics, finance, geography, meteorology, political science, and sociology);
• Use the Box-Jenkins approach to model time series data empirically;
• Run and interpret time-series models involving time-varying (dynamic) volatility and/or trends;
• Use multivariate time-series models such as vector autoregression (VAR) to analyse time series data and apply techniques such as impulse response analysis, Granger Causality, and variance decomposition to interpret results;
• Analyze and interpret cointegrated data in various contexts (such as economics and finance) using an error correction framework, such as the Engle-Granger methodology and the vector error correction (VEC) model;
• Develop fundamental research skills (such as data collection, data processing, and model estimation and interpretation) in applied time series analysis.

[1] Assignments: 30% - 45%

[2] Final Examination: 55% - 70%

Prerequisites for the course include a solid understanding of the fundamentals of probability, statistical inference, regression analysis, and matrix algebra. To be eligible and do well for this course, you need at least an average of 65% (or equivalent) in the final two years of an Australian undergraduate degree with at least two years of university-level statistical and mathematical study including calculus and linear algebra, as well as mathematical statistics and linear regression theory.

Please refer to Course Website