Metadata-Version: 2.1
Name: ndStepwise
Version: 1.0.2
Summary: A package to perform ND Stepwise regression for multiclass problems.
Author-email: Maxwell Dix-Matthews <maxdixmatthews@gmail.com>
Project-URL: Homepage, https://github.com/maxdixmatthews/multiclass-regression
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: OS Independent
Description-Content-Type: text/markdown
Requires-Dist: matplotlib==3.7.4
Requires-Dist: numpy===1.26.0
Requires-Dist: pandas==1.5.2
Requires-Dist: plotly==4.3.0
Requires-Dist: pydot==2.0.0
Requires-Dist: scikit-optimize==0.10.2
Requires-Dist: scikit-plot==0.3.7
Requires-Dist: scikit-learn==1.5.2
Requires-Dist: virtualenv==20.4.3
Requires-Dist: wheel==0.45.1
Requires-Dist: ucimlrepo==0.0.7
Requires-Dist: tensorflow==2.18.0
Requires-Dist: xgboost==2.1.3
Requires-Dist: scipy==1.11.4
Requires-Dist: networkx==2.8.8
Requires-Dist: graphviz==0.20.3

# multiclass-regression

Maxwell Dix-Matthews honours project in multicategory regression

TODO Project:
  1. Add hyperparameter tuning with the digits dataset - this would be a proper case study
  2. Run Kfolds for all datasets (5 results for ND and 5 result for other, try with multiple models too - this may make it more stable as it's got more options?)
  3. Look for more datasets to run it with

TODO Code: 
  1. Look into R's official implementation of ND traversal
  2. Move the cutoff function from model_functions.py to model.py
  3. Make it possible to call a model in the _exact_ same way as scikit
  4. Performance testing with and without threading
  5. Add unit tests
  6. Upgrade to python 3.14 to avoid GIL
  7. Add proper documentation around functions
  
