Title: Parking on random spanning tree in a Hilbert space with applications to tree-based machine learning Authors:  Andrej Srakar - Institute for Economic Research Ljubljana (Slovenia) [presenting]
Abstract: Parking problems on trees have found interesting probability applications in recent years. Yet, many research questions still lack adequate research. The aim is to study the parking problem for a random spanning tree, spreading in a Hilbert space. We will develop scaling limits, large deviations for such a random process, and appropriate limit theorems for the behaviour of parking components. We will point to possible extensions to frozen Erdos-Renyi random graph processes and more general forms of Galton-Watson trees. In application, we will translate and use the theoretical findings for modelling in machine learning using tree-based models, such as random forest or ensemble methods.