Clustering algorithms

[maxkfranz]

Background

Cytoscape.js is a JavaScript library for graph theory analysis and visualisation. It supports many different graph types, such as simple graphs, multigraphs, compound graphs, etc. Cytoscape.js is used in several apps to provide an interactive graph visualisation, and it has also been used for analysis in the terminal and on the server.

Goal

Implement several clustering algorithms for Cytoscape.js.

Description

There are many types of clustering algorithms in graph theory. You would be implementing extensions for the following algorithms using the existing implementations provided by the linked external libraries, creating nice Cytoscape.js APIs:

• Hierarchical : clusterfck
• K-means : clusterfck or maybe kmeans-js
• Affinity propagation : affinity-propagation
• MCL: js-mcl

Depending on time constraints, a number of other algorithms could be implemented:

• K-mediod
• AutoSOME
• Cluster “Fuzzifier"
• Community Clustering (GLay)
• Fuzzy C-Means Clustering (https://github.com/jfinken/fuzzy-cmeans https://github.com/shingt/FuzzyCMeans https://github.com/imrealashu/kmean_fcm)
• Transitivity clustering
• SCPS (Spectral Clustering of Protein Sequences)

  Potential Mentors

Max Franz

Contact

maxkfranz@gmail.com

[devkhan]

I want to learn more about this project. I am familiar with graph theory and some JS graph libraries but not much with Cytoscape.js. Can you please guide me towards some links, resources, etc.?

[maxkfranz]

@devkhan Everything is linked in the post. If you have specific questions, let us know.

[khanspers]

Chosen as GSoC 2016 project.