Goals for 10/26

[ebridge2]

@ebridge2

• [x] begin R package for random timeseries generation, complete with Roxygen2 documentation. Deliverable: package
• [x] read MGC paper. Deliverable: Notes, summary. notes
• [x] is the QC package "general" for other data modalities? http://viz.neurodata.io/freeman14/. Deliverable: plot of calcium imaging data overlap figs compared to mean image new issue
• [x] super simple example that leads to failure. Discussed goals with jovo. Find something that breaks. Have some examples of failures in rds files, and I think I know why they break; to discuss further at office hours tomorrow. example

next week:

• [ ] make FNGS work with BIDs-compliant directories to sweep.
• [x] slice timing correction.
• [ ] FNGS docker.
• [x] discuss ML final project (PCA and CompCor).

@02agarwalt

• [x] Do code.md checklist for EM for GMM and Kmeans with ewok

@ewalke31

• [x] Do code.md checklist for EM for GMM and Kmeans with tanay
  http://neurodatadesign.github.io/fngs/ewok_weekly/EM_GMM.html

[ebridge2]

@02agarwalt see mathematical monk's overview of GMM. Is a very concise, effective overview.

[gkiar]

@ewalke31 where are yours? @ebridge2 @02agarwalt 👌

[jovo]

• [ ] randomts: demo, not description
• [ ] calcium imaging: try using RESTful call directly
• [ ] tGMM: convert to log when having zeros
• [ ] EM_GMM: didn't quite follow code.md afaict. why? failure metric?