Course teached as: B024217 - STATISTICAL INDICATORS: THEORY AND METHODOLOGY Second Cycle Degree in STATISTICS, ACTUARIAL AND FINANCIAL SCIENCE Curriculum STATISTICO
Teaching Language
English
Course Content
In the last years, Social Network Analysis (SNA) has become an active field of research in the statistical framework due the increasing demand of information on a society which is progressively getting more and more "connected". Interesting applications entail, for instance, social integration, import/export flows, migration, etc... The course aims at studying some of the most recent and advanced methodologies in the SNA field.
Textbooks will be suggested at the beginning of the course
Learning Objectives
The course deals with statistical methods and tools for the analysis of social network data. In the first part of the course, emphasis will be given to the study of descriptive tools of analysis. These entail both network visualization and the building of social indicators for the analyses of the structural importance and the cohesion of network's nodes. The second part of the course is devoted to the development of statistical models for social network data. At the end of the course, students will be able to (a) describe and summarize network characteristics; (b) describe and explain objectives and assumptions underlying the different techniques addressed during the course; (c) critically reflect on the methodological adequacy of the modelling alternatives presented during the course.
Prerequisites
Basic knowledge of probability, statistics, and regression modeling
Teaching Methods
Theoretical classes and lab activities
Type of Assessment
The exam consists of a project work and an oral discussion
Course program
- Introduction to social network data
- Network representation: types of relations, graph representation, matrix representation
- Network visualization
- Descriptive analysis of network data: global indicators
- Descriptive analysis of network data: nodal indicators
- Random graph models
- Stochastic blockmodels
- Exponential Random Graph models