Used principle component analysis to classify a data set of 200 galaxies into four categories based on their shape.
The process involved multiple steps including creating the training data set (including the labels) and preparing both data sets for processing.
Six eigenvectors and values were created after processing the data in order to create covariance matrices:
These were then projected onto the base image and then run through a nearest neighbor function in order to tag them as one of the galaxy types.
A more detailed report can be found here.
Team: Aditi Vinod & Lillian Shoemaker
