Bivariate Smoothing Splines

  • Context: Graduate 
  • Thread starter Thread starter Joe Prendergast
  • Start date Start date
  • Tags Tags
    Statistics
Click For Summary
SUMMARY

The forum discussion centers on the need for a bivariate smoothing spline package that allows customization of the loss function. Users have noted that existing tools, such as SCIPY, primarily minimize the sum of squared errors, which does not meet their specific requirements. The discussion highlights the desire to maximize the log-likelihood of a Cauchy distribution of errors instead. Participants suggest using black box optimizers to set up custom problems when existing packages do not suffice.

PREREQUISITES
  • Understanding of bivariate smoothing splines
  • Familiarity with loss functions in statistical modeling
  • Knowledge of SCIPY for spline implementation
  • Experience with black box optimization techniques
NEXT STEPS
  • Research custom loss functions in statistical modeling
  • Explore advanced features of SCIPY for spline fitting
  • Learn about black box optimization methods and tools
  • Investigate alternative packages for bivariate smoothing splines
USEFUL FOR

Data scientists, statisticians, and researchers looking to implement custom loss functions in bivariate smoothing spline models.

Joe Prendergast
Messages
20
Reaction score
4
Does anyone know of a bivariate smoothing spline package that lets you set your own loss function? All of the public domain software I've been able to find (e.g., SCIPY) appears to minimize the sum of squared errors. For example, I'd like to set the spline coefficients to maximize the log-likelihood of a Cauchy distribution of the errors.
 
Physics news on Phys.org
You could just use a black box optimizer and set the problem up yourself. Depending on how many knots you have...
 

Similar threads

  • · Replies 16 ·
Replies
16
Views
2K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 13 ·
Replies
13
Views
3K
  • · Replies 62 ·
3
Replies
62
Views
11K
  • · Replies 10 ·
Replies
10
Views
5K
  • · Replies 4 ·
Replies
4
Views
5K
  • · Replies 67 ·
3
Replies
67
Views
15K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 13 ·
Replies
13
Views
10K