srlearn
srlearn copied to clipboard
srlearn
☕ A Python library for gradient-boosted statistical relational models / learning probabilistic relational programs.
This is an ongoing list of algorithms and learning methods that may be worth exploring/incorporating into `srlearn`. **Classification**: - TILDE (tree learning without boosting) - BoostedMLN - [starling](https://starling.utdallas.edu/software/boostsrl/wiki/mln-boost/), [paper](https://ftp.cs.wisc.edu/machine-learning/shavlik-group/khot.icdm11.pdf) -...
**Describe the bug** Printing the deprecated `BoostedRDN` objects raises a `ValueError`, but printing a `BoostedRDNClassifier` does not. ```python Traceback (most recent call last): File "bug_report.py", line 25, in print(clf) File...
This fails a lot of the tests currently. The idea is to split out the specific solvers from the classifier implementation. Further down the path of "Make BoostSRL optional."
**Describe if the Java version of BoostSRL already supports this.** They're supported in the background format, but they're an undocumented feature that are sometimes mentioned in examples. **Is your feature...
**Describe if the Java version of BoostSRL already supports this.** Usually it's assumed that the user manually splits the data into folds or standard folds exist (e.g. https://github.com/srlearn/datasets/) - There...
**Describe if the Java version of BoostSRL already supports this.** Yes. As described [here](https://starling.utdallas.edu/software/boostsrl/wiki/advice/), BoostSRL supports learning with advice. **Is your feature request related to a problem? Please describe.** No...
Probably need to think on this a bit. - [GitHub CITATION files docs](https://docs.github.com/en/github/creating-cloning-and-archiving-repositories/creating-a-repository-on-github/about-citation-files) - [ruby-cff](https://github.com/citation-file-format/ruby-cff) - [citation-file-format](https://github.com/citation-file-format/citation-file-format)
**Describe the bug** It looks like running a notebook cell with `RDN.fit()` multiple times doesn't cleanup previous state. During interaction: ```console >>> from srlearn.system_manager import reset >>> reset(soft=True) ['data5', 'data9',...
Line 43: Phase out this function Line 168: treeDepth vs. maxTreeDepth