Publication Date

Spring 2019

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science

First Advisor

Katerina Potika

Second Advisor

Christopher Pollett

Third Advisor

Sami Khuri


Community detection, Local graph clustering, Online com- munity


A community in a network is a group of nodes that are densely and closely connected to each other, get sparsely connected to the nodes outside the community. Finding communities in a large network helps solve many real-world problems. But detecting such communities in a complex network by focusing on the whole network is not feasible. Instead, we focus on finding communities around one or more seed node(s) of interest. Therefore, in this project, we find local communities. Moreover, we consider the online setting where the whole graph is unknown in the beginning and we get a stream of edges, i.e., pair of nodes, or a stream of higher order structures, i.e., triangles of nodes.

We created a new dataset that consists of web pages and their links by using the Internet Archive. We extended an existing online local graph community detection algorithm, called COEUS, for higher order structures such as triangles of nodes. We provide experimental results and comparison of the existing method and our proposed method using two public datasets, the Amazon and the DBLP as well as for our new Webpages dataset. In the experimental results, we see that the proposed method performs better than the existing method for one out of three test cases for the public dataset but not for our Webpages dataset. This is because the Webpages dataset has a large number of nodes with degree 1 which poses a problem for modified COEUS because it takes triangles as an input stream.