A really nice TE day, even with a lot of rain

Tasks

• Pay Vanessa

SBIRs

• 9:00 Standup
• Working on making cluster datasets – done. Took the best looking option and reduced it to 2, 3, 5, 8, 13 dimensions. We’ll try clustering on those to see what the ratios of clusters to unclustered are. Then maybe take the best number of dimensions and apply that to the best result?
  • That really didn’t do much. The clusters just got smaller as the dimensions went up. Which, in retrospect, is unsurprising.
  • What I really should do is chain the UMAP to the HDBSCAN and do a parameter space exploration that determines optimal UMAP parameters per dimension.
• 4:00 ADS meeting – We are now looking at a Phase III? And still no progress on the NDS follow-on. We will need to write up a description of the plan for initial integration after the meeting with Matt next week