Research

🔬 My Research Focus

My research focuses on clustering longitudinal mixed-type data for modeling longitudinal surveys. I develop statistical methods that can handle the complexity of real-world data that evolves over time and contains different types of variables (continuous, ordinal, categorical). My work has applications in healthcare, social sciences, and survey research, particularly in understanding behavioral patterns and policy impacts over time.

I completed my PhD in Applied Mathematics at Université Lyon 2 under the supervision of Julien Jacques and Isabelle Prim-Allaz, with my thesis defended in May 2025.

🔍 Statistical Clustering

Finite mixture models for complex data structures

📊 Longitudinal Data

Time series analysis and temporal pattern recognition

🔀 Mixed-Type Data

Methods for heterogeneous data types

📚 Google Scholar 🔬 HAL Profile 🆔 ORCID

📖 Publications

Journal Articles

A Comparison of Migrant Integration Policies via Mixture of Matrix-Normals

Salvatore Leonardo Alaimo, Francesco Amato, Filomena Maggino, Alfonso Piscitelli, Emiliano Seri

Social Indicators Research, vol. 165, pp. 473–494 (2023)

DOI HAL Published

Clustering longitudinal ordinal data via finite mixture of matrix-variate distributions

Francesco Amato, Julien Jacques, Isabelle Prim-Allaz

Statistics and Computing, vol. 34, article 81 (2024)

DOI HAL Published

Under Review

MMM: Clustering Multivariate Longitudinal Mixed-type Data

Francesco Amato, Julien Jacques

Submitted to Computational Statistics and Data Analysis (2025)

HAL Preprint Under Review

PhD Thesis

Mixed data temporal clustering for modelling longitudinal surveys

Francesco Amato

PhD Thesis, Université Lumière Lyon 2 (Expected May 2025)

Supervisors: Julien Jacques, Isabelle Prim-Allaz
Laboratory: ERIC (Entrepôts, Représentation et Ingénierie des Connaissances)

DOI HAL Defended

🔬 Research Projects in Progress

Model-based Clustering for Longitudinal Mixed-type Data: a Survey

Ongoing Drafting

Literature survey on clustering methods for longitudinal, mixed-type and longitudinal mixed-type data, based on Chapter 3 (state of the art) of my PhD thesis.

MMMClust

Ongoing Implementation

R package implementing the MMM algorithm for clustering multivariate longitudinal mixed-type data. It will include comprehensive documentation, examples, and validation datasets.

🎯 Research Focus

🔧 Statistical Methodology

Development of finite mixture models specifically designed for longitudinal data with mixed variable types (continuous, ordinal, categorical). Focus on matrix-variate distributions and latent variable approaches.

📐 Mathematical Framework

Advanced probability theory, optimization algorithms (EM algorithm variations), and matrix calculus for handling complex data structures that evolve over time.

💡 Applied Research

Real-world applications in longitudinal surveys, social policy analysis, and financial data clustering with focus on practical implementation and interpretation.

💻 Software & Code

Research Code Repository

Reproducible code for published research papers with README files, sample datasets, and step-by-step analysis scripts available on my GitHub page. Some open, some available on request for collaboration purposes.