More Ground-Truth Testing
I just completed additional testing on the ground-truth graphs. Specifically, I’ve calculated the metric scores on output graphs in order to determine if the Louvain-{x metric} variant of the algorithm actually outputs the clustering with the best x metric score. For example,…
Running the Data Set
Our data set consists of 52 networks. The size of the networks range from 34 vertices and 78 edges (Zachary’s karate club network) to 4,847,571 vertices and 68,993,773 edges (LiveJournal social network). We ran two tests on our data set:…
Louvain Method Based on the Silhouette Index Metric
Anastasia and I recently finished implementing the Louvain algorithm based on the silhouette index (SI) metric. This post discusses our process. For a full implementation of any metric in the Louvain algorithm, the algorithm must be able to calculate the…
Creating the Adjacency Matrix
We are now able to construct an adjacency matrix of the network. As Anastasia explained in a previous post, we will use the adjacency matrix in silhouette index to find the shortest path between all pairs of vertices. Our matrix…
Input/Output Issue
The input file represents the network. It should be formatted as a space-delimited file, where each line contains the start node, destination node, and an optional edge weight. The test network. It is undirected and unweighted. The input file for…
Switch to Java
We recently switched from the C++ to the Java implementation of the Louvain method. Our main concern with the C++ implementation was how it handled serialization. With C++, we had to deal with the serialized version of the network to…
Quality Metrics
Community quality metrics are measurements of how well communities are formed. Good metric values generally reflect networks where connections are denser within communities than between communities. We are getting ready to implement various quality metrics in the Louvain method, an…
Summer Research
This past summer I was a research assistant for the computer science department at Pomona College through the Summer Undergraduate Research Program (SURP). I worked with Professor Chen on a project involving community-detection algorithms for complex networks. I looked at…