USI - Email
Inference for dynamic systems: application to infectious diseases
Host: Prof. Ernst Wit
USI Lugano Campus, room SI-003, Informatics building
University of Haifa, Israel
Understanding multi-group (age-structure/spatial-structure) dynamics of infectious diseases is a fundamental issue for both scientific study and policy making. In particular, the question of how to define a data driven criteria for partitioning incidence data to specific groups has yet to been addressed. In this talk we present a statistical methodology for data-driven partitioning of infectious disease incidence into groups, and fitting a model to the resulting partition. Specifically, we present a methodology for building a hierarchical clustering tree and pruning it, leaving only significant age-group clusters in the partition, while controlling the familywise error rate. The partition algorithm is built on recent theoretical results on identifiability of parameters underlying the mechanistic models of epidemic dynamics. The methodology is tested using simulations and applied to infectious disease incidence data of seasonal influenza.
Itai Dattner is on the faculty of the Department of Statistics at the University of Haifa since 2013. He received his Ph.D and M.A. degrees in Statistics from University of Haifa in 2011 and 2007 respectively. He was a post-doctoral researcher at EURANDOM, Department of Mathematics and Computer Science, Eindhoven University of Technology, The Netherlands, where he was involved in joint research project with Philips Research Eindhoven, Department of Molecular Diagnostics. While doing the PhD, he served as Data Manager and Biostatistician in the EGeR Clinical Trial. Itai Dattner's research is supported by the German-Israeli Foundation (The GIF Young Scientists’ Program), and the Israeli Science Foundation (ISF). Itai Dattner serves as the Israeli representative for the Executive Committee of the Eastern Mediterranean Region (EMR) of the of the International Biometrics Society (IBS). He is doing consulting work in a variety of fields such as digital advertising, mobile networks, finance and big data companies.
INF Newsletter Archive