Version 1.1.0
What's Changed
This release adds the possibility of using pandas dataframe as input and output for inserting regression models. Those dataframes may contain columns of Gurobi variables or constants (fixed features). This is particularly convenient when used in conjunction with gurobipy-pandas.
We also add the possibility of handling Scikit Learn column transformers. In conjunction with pandas input, this makes it much more easier to handle variables that are indexed by categorical features.
Those two features are illustrated in the student enrollment example and the price optimization example.
This release also introduces the ability to use Scikit Learn PLS Regression. Thanks to @DavidWalz for contributing it!
The formulation of the decision tree has also been improved so that if should be faster to generate the models.
Finally, the documentation has been updated to include summary explanations on the MIP formulations used to represent the various regression models, the potential sources of differences with the original regression models and how to remedy them. The new page can be found here.
Relevant pull requests
- Fixed torch notebook to use PyTorch by @Epanemu in #99
- add PLS Regression by @DavidWalz in #98
- Faster decision trees by @pobonomo in #113
- Alternative implementation for integrating with pandas by @pobonomo in #111
- More docs by @pobonomo in #120
New Contributors
- @Epanemu made their first contribution in #99
- @DavidWalz made their first contribution in #98
Full Changelog: v1.0.1...v1.1.0
