search

h2oai / h2o-3

H2O is an Open Source, Distributed, Fast & Scalable Machine Learning Platform: Deep Learning, Gradient Boosting (GBM) & XGBoost, Random Forest, Generalized Linear Modeling (GLM with Elastic Net), K-Means, PCA, Generalized Additive Models (GAM), RuleFit, Support Vector Machine (SVM), Stacked Ensembles, Automatic Machine Learning (AutoML), etc.
http://h2o.ai
Apache License 2.0
6.92k stars 2k forks source link

Explore if GAM can run without any knots and not, why not #6925

Closed exalate-issue-sync[bot] closed 1 year ago

exalate-issue-sync[bot] commented 1 year ago

Wendy Wong commented: # For bs = 0, cubic spline, minimum number of knots is 3, otherwise there will be problem with matrix inversion and the code won’t run;

For bs = 1, thin-plate, minimum number of knots is M+1 where M is generated from some formula. I have no control here on how k must be set;

For bs = 2, minimum number of knots is N=2. When you you set N=2, the number of basis function will change depending on the order you choose for the splines:

   for order = 1, there will be one basis function;

   for order = 2, there will be two basis function as explained here:

GAM always need knots, however, for I-splines, it can have only one basis function if you set number of knots = 2 and order = 1.

exalate-issue-sync[bot] commented 1 year ago

Wendy Wong commented: The second part of this Jira will be to implement support when number of knots < spline_degree.

h2o-ops commented 1 year ago

JIRA Issue Details

Jira Issue: PUBDEV-8836 Assignee: Wendy Wong Reporter: Wendy Wong State: Resolved Fix Version: 3.38.0.3 Attachments: N/A Development PRs: Available

h2o-ops commented 1 year ago

Linked PRs from JIRA

https://github.com/h2oai/h2o-3/pull/6333 https://github.com/h2oai/h2oai-serving/pull/1112 https://github.com/h2oai/h2oai-serving/pull/1135