search

nrnb / GoogleSummerOfCode

Main documentation site for NRNB GSoC project ideas and resources
114 stars 38 forks source link

Implement Principal Coordinate Analysis in clusterMaker #45

Closed scootermorris closed 8 years ago

scootermorris commented 8 years ago

Background

One technique for dimensionality reduction is principal coordinate analysis (PCA). This technique is an analog of principal component analysis, but it works on distance matrices, which are not appropriate for principal component analysis. Adding this to clusterMaker would allow users to select the edges to retain based on an analysis of the contributed variance.

Goal

The goal of this project is to add PCA to clusterMaker and provide useful visualizations that will allow users to select the components to remove from the network and those to retain. As networks get larger and larger, providing tools to reduce the dimensionality of graphs is going to be increasingly important, and this could provide one approach to doing so.

Difficulty level: 1

Prior experience with Cytoscape app development is not required.

Technology and Skills

Java, Cytoscape, matrix math

Potential Mentors

Scooter Morris

Contact

scooter@cgl.ucsf.edu

SupunArunoda commented 8 years ago

Are there any errors in github example repository? Some exceptions have occurred while run them using maven...

khanspers commented 8 years ago

Hi. Which repository are you referring to? If you have questions about the PCA in clustermaker project, please contact the mentor directly by means of the recommended info (see "Contact" above). They might not see the questions here.

khanspers commented 8 years ago

Chosen as GSoC 2016 project.