April schedule
Friday 14 April, 4 PM
Anna Aksamit, Senior Lecturer, School of Mathematics and Statistics, University of Sydney
Hybrid: Red Centre 4082
https://unsw.zoom.us/j/88495626621
“Modelling an additional information: mathematical tools and financial applications”
Abstract: In this talk I will discuss the theory of enlargement of filtration which has been used in modelling asymmetric information. I will review classical results and present some new development with financial applications. My main focus will be semimartingales calculus in different filtrations.
Friday 21 April, 4 PM
Michael Lydeamore, Lecturer, Econometrics & Business Statistics, Monash University
Virtual
https://unsw.zoom.us/j/88495626621
“Data-Driven Insights into Healthcare Challenges: Two Case Studies”
Abstract: In this seminar, I will present two data-driven healthcare projects. The first project aims to calculate the burden of healthcare-acquired infections in Australia. We conducted a point prevalence survey for five such infections in 2018 and compared their burden with that of more recognized conditions requiring hospitalization. We found that these largely preventable conditions lead to almost 7,000 deaths annually, highlighting the need for significant improvement in healthcare practices.The second project focuses on understanding the spread of antimicrobial resistance (AMR) in Victorian hospitals. Using linked data from the Victorian State Government, we analyzed patient movement across the healthcare network to design surveillance and control strategies in case of an outbreak. With 8 million hospital admissions per year, it would be ideal to test strategies on a small subset of this network. However, it is unclear whether AMR-infected patients move around the network in the same way as non-infected patients. To address this, we used patient matching and survival analysis and found that AMR-infected patients move around the network to the same number of places as AMR-negative patients, but at a faster pace. This highlights the critical need for timely control strategies to contain the outbreak.
Friday 28 April, 4 PM
Kassell Hinge, Postdoctoral Fellow, College of Business and Economics, Australian National University
Virtual
https://unsw.zoom.us/j/88495626621
“Rapid Implementation of Score Matching Estimators”
Abstract: Score matching (Hyvarinen, 2005) is an estimation technique that avoids normalising constants in model densities (i.e. it works on improper densities) and can thus be used in many cases that maximum likelihood estimation cannot. Unfortunately, the implementation of score matching estimators often requires tedious calculus, especially for models of data that lie on multidimensional manifolds. I have developed an R package that uses automatic differentiation to make implementation much faster. The package already contains estimators for compositional data models and directional distributions.