Questions and remarks
- Is incremental fitting a good approach? Burr Settles is not clear on this one.
- Active learning in sparse situations. Always labeling 0 is the best approach. How are weights added to the model. How are these computed? From the posterior?
Committing guidelines:
- Commit machine-readable and FOSS formats as much as possible
- repect dir structure, use the Rstudio project, Makefile etc.
- Don't commit huge datafiles (>20MB)
- Communicate on issue tracker and close issues from commits
- No spaces in filenames please
- Python scripts in PEP8 and YAPF formatted