Proposal Description

The increasing availability of heterogeneous, high-dimensional, and structurally complex data—ranging from distributional and interval-valued observations to spatial, temporal, and manifold-based structures—demands advanced statistical methodologies capable of modeling, synthesizing, and interpreting nonstandard data forms. This session brings together researchers contributing innovative approaches rooted in Bayesian modeling, information-theoretic inference, clustering for symbolic and spatial data, interpretability techniques for machine learning, and dimension reduction for complex data spaces.
The invited contributions illustrate recent methodological advances and their practical relevance across environmental, health, and socioeconomic applications. Topics include Bayesian spatial modeling for misaligned environmental and epidemiological data, clustering methods for spatial interval time series, explainability techniques tailored to modern high-dimensional predictive models, and information-theoretic frameworks for inference under partial identification and complex data structures.
The session aims to foster interdisciplinary dialogue and stimulate new research directions in the statistical modeling of complex, symbolic, and distributional data, highlighting the crucial role these methodologies play in modern Statistics and Data Science.