Statistical methods for large, complex and structured data

The research project will focus on the development of novel statistical methodology for inference in large, complex and structured problems – such as those arising from high and ultra-high dimensional data; graph and network-valued data; functional data; multi-modal data; etc. A particular emphasis will be placed on sparse and interpretable statistical learning methodology; scalable techniques for feature selection, dimension reduction and assumption-free testing; computational assessments of significance and stability; and approaches for latent structure recovery and clustering. The research will take place within the interdisciplinary environment supported by the L’EMbeDS Department of Excellence (Economics, Management and Law in the era of Data Science), drawing upon statistical theory and computational tools to foster multiple collaborative endeavors and translate methodological innovation into impactful domain applications.