Nowadays, large amounts of data describe how entities interact with each other. For example, these data may represent friendship relations, co-working interactions between colleagues, financial transfers between banking institutions, or functional connectivity between different areas of the brain. Network data are the mathematical tools that are most frequently used to represent and store these interactions. Researchers are often interested in modeling such network data, with purposes including the understanding of how interactions are created, the identification of what factors make two units more likely to interact, and the detection and description of some features of interest.
Statistical network analysis encompasses methods that account for the complex dependencies found in network data, with the goal of data modeling and description, dimensionality reduction, and identification of interesting patterns. This short course aims at giving an introduction to some of the main methods for the analysis of network data, with particular focus in model-based data analysis.
Tentative content below:
1. Network data
2. Introduction to graph theory
3. Models for random graphs
4. Clustering and block models
5. Latent space models
6. Other topics
Hands on tutorials using the statistical software R will be used to present the methods and showcase their application to real-world data (it is recommended that packages “igraph”, “igraphdata”, “networkdata”, “blockmodels”, “latentnet”, “huge”, and “covglasso” are installed).
References:
- Kolaczyk , Csárdi (2020) - Statistical Analysis of Network Data with R
https://doi.org/10.1007/978-3-030-44129-6
- Salter-Townshend et al (2012) - Review of statistical network analysis: models, algorithms, and software. Statistical Analysis and Data Mining
https://doi.org/10.1002/sam.11146
Calendario del corso
- Lunedì 8 maggio, ore 11-13 e 14-16, aula N Belmeloro
- Venerdì 12 maggio, ore 9-11, lab G Ranzani
La partecipazione è gratuita, il corso è aperto agli studenti delle Lauree Magistrali Internazionali e ai Dottorandi del primo anno, ed occorre registrarsi a questa pagina entro il 30 aprile 2023.
Qualora si fosse interessati a partecipare, ma non si è iscritti ai suddetti corsi, scrivere a laura.anderlucci@unibo.it