Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

best_error sometimes get trapped at 0.0000 #38

Closed
cylammarco opened this issue Jun 23, 2021 · 3 comments
Closed

best_error sometimes get trapped at 0.0000 #38

cylammarco opened this issue Jun 23, 2021 · 3 comments
Assignees
@jveitchmichaelis
Labels
enhancement New feature or request urgent
Projects
v0.3.0
Milestone
V0.3.0

Comments

@cylammarco
Copy link
Collaborator

During fit, the best_error sometimes gets stuck at 0.0000 with the number of best inliners equal to sample_size, which is clearly a bad over-fit. We should have some defensive statement to avoid "perfect fit" to overtake good fit with high peak utilisation fraction.
@cylammarco cylammarco added enhancement New feature or request urgent labels Jun 23, 2021
@cylammarco cylammarco added this to the V0.3.0 milestone Jun 23, 2021
@cylammarco cylammarco added this to To do in v0.3.0 via automation Jun 23, 2021
cylammarco added a commit that referenced this issue Jun 24, 2021
@cylammarco
Fixed #42 by checking for finite and non-nan values. Potentially fixed …
b974dbe 
…#38 by acception best_err only if the error is larger than 1e-4.
@cylammarco
Copy link
Collaborator Author

Now set to accept best_err only if the error is larger than 1e-4, keep open as this condition may not be sufficient.

jveitchmichaelis added a commit that referenced this issue Jul 4, 2021
@jveitchmichaelis
add options for minimum matches/peak utilisation. Addresses #38
e85efa7
@jveitchmichaelis
Copy link
Owner

I think this is now addressed with the referenced commit, but please have a look.

@cylammarco
Copy link
Collaborator Author

This doesn't seem to be a problem anymore, close for now.

@cylammarco cylammarco moved this from To do to Done in v0.3.0 Aug 12, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@jveitchmichaelis jveitchmichaelis

Labels
enhancement New feature or request urgent
Projects
No open projects
v0.3.0
  
Done
Milestone
V0.3.0
Development

No branches or pull requests
2 participants
@cylammarco @jveitchmichaelis