cobra
cobra copied to clipboard
PythonPredictions
A Python package to build predictive linear and logistic regression models focused on performance and interpretation
# Add functionality to evaluate bias in modelling outputs Task: Adding functionality to model evaluation code to generate 1 or multiple plots to assess if the model predictions are biased...
# 143 serialization deserialization bug [143 serialization deserialization bug] (https://github.com/PythonPredictions/cobra/issues/143) ## Changes made Added check in target_encoder.py to include multiple types of floating numbers. ## How does the solution address...
# Bug Report After serializing and de-serializing a PreProcessor with only contiguous variables (to check if it is also the case when categorical variables are present) 1. the preprocessor object...
Resolves #171 This PR refactors the existing `compute_univariate_preselection` method according to the DIY principle. It also prepares for sending a custom scoring metric.
This method has a responsibility for classification and for regression. This can be simplified, since the regression part contains a lot of code from the classification part.
Visualize contribution of features to a single prediction (or average contribution for multiple predictions)
Add functionality for easy plotting of profiling & profit graphs
Add at least 1 more complex benchmarking model, or facilitate integrating external models into a comparison
We compute variable importance by calculating Pearson's correlation between scores and target encoded variables: https://github.com/PythonPredictions/cobra/blob/0133435194ac63f771e6d165c23ea34233dd0a94/cobra/model_building/models.py#L144-L150 It'd be nice to choose different correlation (like Kendall)? Pearson assumes normality, but doesn't always...
When the user builds multiple different models, it would be handy to evaluate them in the same time. This means that the plotting functions: - evaluator.plot_roc_curve() - evaluator.plot_confusion_matrix() - evaluator.plot_cumulative_gains()...
Metadata
Owner
Metadata
A Python package to build predictive linear and logistic regression models focused on performance and interpretation