Synopsis: Students read and collaboratively discuss recent research papers in multivariate dynamic modelling for time series and forecasting, linked to current research frontiers, and motivating applications in areas of business, IT, economics and allied areas. Class discussion are led by students presenting overviews/reviews and their individual work on exploring models, relevant statistical theory and methods. Students may build on this seminar to develop future extensions (theoretically and computationally) and applications in new areas.

Prerequisites: STA 831 and 642 (or instructor permission).

Full preparation and background in all material in the texts below is assumed.

• Time Series: Modeling, Computation, and Inference, by Raquel Prado & Mike West, 2010, Chapman Hall/CRC Press Taylor & Francis Group.
• Bayesian Forecasting & Dynamic Models, by Mike West & Jeff Harrison, 1997 (2nd edition), Springer-Verlag.

Registration: Instructor permission required. Registration restricted to PhD students in Statistical Science

Graduate credit units: 1

Assessment/Grading: P/F
Full participation and engagement: (a) presentations and discussion-leading from each student in topic-specific meetings, leading discussions based on reading of research papers and their individual work on topics in the papers; and (b) every student engages and contributes in discussion and Q&A on all topics presented by other students.

Reading: Recent research literature linked to the seminar page for registered students.

---

Duke Graduate School Community standards

Duke Green Classroom certified