Projects

For each project write the report and add the files with programs and constructed data sets (except for very large data). Pack everything in a ZIP file. Send the ZIPs for Project 1 and Project 2 to vladimir.batagelj@fmf.uni-lj.si, and for Project 3 to luka.kronegger@fdv.uni-lj.si till September 18th, 2022.

Solve the problem with your number with the Monte Carlo method. If you are able to solve it also theoretically compare both results.

Do either the Task A:

Transform the data from a selected data source into a Markov chain transition matrix with at least 15 states and analyze it. For example, the transition matrix from a given text (project Gutenberg) that is determined by consecutive letter pairs.

or Task B:

On a labeled network of your choice with at least 15 nodes and a measured node property (attribute) either

• Compute Geary's and Moran's measure.

or for a selected network property

• Make a CUG or QAP test.

Interpret the obtained results.

You will work on data that you choose from the list of networks available in Statnet Data. Find the panel data measured in at least two waves (you don't need to choose the first two). Your task is to build ERGM for the first chosen wave and SAOM for the two waves. Models have to model clustering and homophily by selected “node level” variable. Then you need to assess the GOF too.

Respond to the following questions:

1. What are the main characteristics of the network in the two observations? Are they similar/different?
2. Which micro patterns are important in explaining the structure of the selected network?
3. Which micro-mechanisms shape the evolution of your selected network?

Snet