Community detection in graphs (Alex Levin, PyData TLV - Oct 21)

Detect communities in graphs using bipartite models, graph theory, and mathematical modeling to analyze complex networks, identify patterns, and detect anomalies.

Key takeaways

Community detection can be achieved by modeling the similarity between nodes and using a bipartite graph.
The affiliation graph model (AGM) can be used to describe overlapping communities.
Modularity is a measure of the quality of community detection, calculated as the difference between the number of edges within a community and the number of edges expected by chance.
The Levene algorithm is a method for detecting non-overlapping communities, but is not suitable for overlapping communities.
The AGM model can be used to generate a graph that is similar to the original graph.
The Louvain algorithm is a method for detecting overlapping communities.
Community detection can be applied to social networks, such as online communities or networks of friends.
Graph theory and mathematical modeling can be used to analyze and understand complex networks.
Gradient descent can be used to optimize the AGM model and find the best community partition.
Modularity is a non-linear measure, and the AGM model can be used to optimize it.
The AGM model can be used to generate a graph that is similar to the original graph, and to detect overlapping communities.
Community detection can be used to identify malicious networks and to detect anomalies in complex systems.
The AGM model is a flexible model that can be used to describe overlapping communities.
The Louvain algorithm is a fast and efficient method for detecting communities, but is not suitable for overlapping communities.
Modularity is a measure of the quality of community detection, and can be used to evaluate the performance of community detection algorithms.
Community detection can be used to identify patterns and structures in complex networks.
The AGM model can be used to generate a graph that is similar to the original graph, and to detect overlapping communities.
Graph theory and mathematical modeling can be used to analyze and understand complex networks.
Community detection can be used to identify malicious networks and to detect anomalies in complex systems.
The Louvain algorithm is a method for detecting non-overlapping communities, but is not suitable for overlapping communities.

Community detection in graphs (Alex Levin, PyData TLV - Oct 21)

More talks