Social-Networks-Analysis

Lab 1 - Analysis & Study of Complex Network Topologies

In the first lab exercise the following are implemented:

1. Construction and visualization of complex types of networks
2. Calculation of network metrics: a) clustering coefficient, b) Minimum path length c) Node eccentricity: diameter, radius, circumference, center
3. Calculating node centrality metrics: degree centrality, closeness centrality, betweenness centrality, Katz centrality, and applying PageRank to a real network - web-Stanford
4. The study of connectivity and robustness of networks
5. The study of evolutionary network conversion
6. The study of real networks - GoT

Lab 2 - Analysis of Social Structure in Artificial and Real Complex Network Topologies

In the second lab exercise the following are implemented:

A. Study of real topologies

• American College Football
• Game of Thrones - season 5
• email-Eu-core & ground-truth division

Node degree distribution and average degree of each topology. Node clustering factor distribution and the average clustering factor of each topology. Closeness centrality distribution and mean closeness centrality of each topology.

B. Familiarity with community detection tools

Using the SpectralClustering, girvan_newman and greedy_modularity_communities functions for each real and synthetic network:

1. Visualize the communities resulting from each method
2. Make the necessary comments (eg: comparison of the number of communities calculated by each algorithm).

Lab 3 - Link Prediction

In the third lab exercise the following are implemented:

1. Graph construction and preprocessing for link prediction
2. Introduction to similarity-based metrics for Link prediction
3. Link prediction based on similarity-based metrics
4. Link prediction with embedding based on random walks (Random Walks)

The Similarities dataset DBpedia is used.