C2. High dimensional, Educational and Functional data

17:20 - 18:30, Aula 10

Chair: Giovanni Boscaino

Analysis of Brain-Body Physiological Rhythm Using Functional Graphical Models

Rita Fici, Luigi Augugliaro and Ernst C. Wit

Abstract: This paper presents an analysis of physiological data derived from a recent investigation on network physiology, adopting the conceptual framework that views the human organism as a complex network of interacting organs. The study explores coordinated interactions among organs using functional conditional Gaussian Graphical Models (fcGGM). Organ functions are modelled as networks with individual regulatory mechanisms, forming a broader system through continuous interactions. The focus of the analysis is on the interactions within and between two subnetworks: brain activity and a composite network comprising the RR interval of the electrocardiographic waveform, respiration amplitude and blood volume pulse.

Click here to view the abstract.

A comparison of scalable estimation methods for large-scale logistic regression models with crossed random effects

Ruggero Bellio and Cristiano Varin

Abstract: Parameter estimation of generalized linear models with crossed random effects for large-scale settings is hampered by challenging numerical hindrances. This contribution focuses on logistic regression with crossed-random intercepts and it investigates the properties of two estimation methods for which a scalable software implementation exists, namely the all-row-column and penalized quasi- likelihood methods. The results of a simulation study for sparse settings inspired by e-commerce data, with sample sizes up to \(10^6\) suggest that the all-row-column method is preferable over penalized quasi-likelihood.

Click here to view the abstract.

Single-cell Sequencing Data: Critical Analysis and Definition of Statistical Models

Antonino Gagliano, Gianluca Sottile, Nicolina Sciaraffa, Claudia Coronnello and Luigi Augugliaro

Abstract: In the past decade, advances in single-cell RNA sequencing technologies have radically improved the comprehension of cell biology. Pseudo-time computation has revealed greater heterogeneity in cell differentiation, which is important for the study of various diseases. In this paper we evaluate pseudo-time calculation methodologies and propose statistical models for analyzing the relationship between pseudo-time and gene expression.

Click here to view the abstract.

Investigating the association between high school outcomes and university enrolment choice: a machine learning approach

Andrea Priulla, Alessandro Albano, Nicoletta D’Angelo, Massimo Attanasio

Abstract: This paper investigates how proficiency in mathematics and Italian tests in high school affect university enrolment choices in Italy. We distinguish between students from scientific and humanistic backgrounds, providing valuable insights into their enrolment choices. We employ gradient boosting methodology, adjusting for students’ sociodemographic characteristics and previous educational attainment. Results shed light on the interplay between student performance, sex, and the type of high school attended in shaping enrolment choices.

Click here to view the abstract.