Issue 856: When training a "Default" peak scoring model, the model should be able to use feature scores that are present for some but not all peptides

issues
Status:closed
Assigned To:Guest
Type:Defect
Area:Skyline
Priority:3
Milestone:22.1
Opened:2021-12-06 by Nick Shulman
Changed:2022-03-30 by Nick Shulman
Resolved:2022-02-15 by Nick Shulman
Resolution:Fixed
Closed:2022-03-30 by Nick Shulman
2021-12-06 Nick Shulman
Title»When training a "Default" peak scoring model, the model should be able to use feature scores that are present for some but not all peptides
Assigned To»Nick Shulman
Type»Defect
Area»Skyline
Priority»3
Milestone»22.1
When you train a Default peak scoring model, it does not end up using the full set of features that the untrained Default model.
The untrained Default scoring model never has missing (or "unknown") scores, since it always substitutes 0 for the missing score value.
 
 MissingScores.png

2022-02-15 Nick Shulman
resolve as Fixed
Statusopen»resolved
Fixed in PR 2001

2022-03-30 Nick Shulman
close
Statusresolved»closed
Assigned ToNick Shulman»Guest