Use graph output from ndgm on Healthy Brain Network MRI data to perform feature learning with Agglomerative Clustering and Gaussian Mixture Models.

[GaneshArvapalli]

Description:

Use graph output from ndmg on Healthy Brain Network MRI data to perform feature learning with various deep learning techniques.

DoD:

• Pipeline that takes ndmg output on MRI graphs stored in S3 and feeds it into graph analysis python code (either custom written, or GrasPy) by Oct. 15th
• Python script for feature reduction/projection using PCA by October 20th
• Output of at least 10 graph features for every edgelist file in S3 in matrix format (stored in a file) by October 25th
• Generalized R/Python script that separates data into clusters (with tests) using GMM and plots output by October 31st
• Generalized R/Python script that separates data into clusters (with tests) using Agglomerative Clustering and plots output by November 6th
• Generalized R/Python script that separates data into clusters (with tests) using Ward's Method and plots output by November 12th
• Formatted data output that can be stored in a file for usage by either R or Python by November 16th (extra time will be used in case of bug fixing or lack of clusters)

[jovo]

can't start with DL

[jovo]

@GaneshArvapalli make sure goals are SMART, which algorithm, an "implementation" is not a deliverable, working tests, or analyzed data, etc. please update.