Week 1
I met with Dr. James Abello, my mentor, to discuss about the research project. I was given some papers to read on the topic beforehand which helped me get a general understanding about the project. Jack and I worked on our first presentation which we have to give on Tuesday, June 7, 2011. I also worked on setting up my Webpage.
Week 2
Jack and I gave our first presentation about our project. I met with my mentor to talk about the progress in the project. I familiarized myself with the graph visualization and analysis tool, Cytoscape, which we will be using to analyze and manipulate graphs. We worked on setting up a XML-RPC connection in Cytoscape by using a plugin called Cytoscape-RPC which allows the user to create, modify and query graphs. The purpose of this connection was to to set up a communication between the visual tool and the clustering algorithm.
Week 3
Jack and I worked on a simple graphical user interface that displays graphs and allows the desired user interactions. I read "Finding Community Structure in Very Large Networks" by Clauset, Newman, and Moore, that described the fast modularity algorithm which we will be using to discover communities in graphs.
Week 4
Dr. Tina Eliassi-Rad sent us a Java implementation of the fast modularity algorithm which included a GUI for importing and visualizing graphs. Jack and I spent some time understanding how this program works. We then worked on adding desired features to this useful tool.
Week 5
Jack and I met with our mentors to discuss about our progress. After familiarizing ourselves with the community analysis tool that our mentor, Dr. Tina Eliassi-Rad provided us with, we started adding some desirable features to this code. Our goal is to bring the human into the loop in the clustering process. Thus, we would like to have some functions that allow the user to manipulate graphs in this tool. We added a function that allows the user to move nodes from one community to another. We had an REU event in which we visited the Cancer Institute in New Brunswick. There we had a tour and then we listened to a few talks from Math and Computer Science researchers in this institute.
Week 6
Jack and I continued to work on the implementation of some desired features in the community analysis tool. Our goal is to get feedback form the user in the clustering process. And thus far the user is able to change the community of a node, create a new community in the graph, and the Modularity which is the metric that we are using to cluster graphs (see Modularity) will be recomputed and displayed to the user.
Week 7
Jack and I started thinking more about computational issues involved in the clustering of graphs. We added other features to the community analysis tool. We also prepared our final presentation which we gave on July 14, 2011.
Week 8
Jack and I wrote our final report about our work this summer.