/Logistic-Regression-with-IRLS

Project for studies course Andvanced Machine Learning.

Primary LanguageJupyter Notebook

Logistic-Regression-with-IRLS

Project for studies course Andvanced Machine Learning.


File structure

There are 3 categories of files here: simulation experiments, experiments on real data, and results (plots).

Simulation experiments:

  • simulation_experiments.ipynb - Jupyter notebook showing the experiments on simulated data sets,
  • simulation_functions.py - functions used in this notebook.

Real data experiments:

  • real_data_experiments.ipynb - Jupyter notebook showing the experiments on 2 data sets,
  • real_data_experiments.py - a Python script that conducts the same experiments as the notebook,
  • data_preprocessing_breast_cancer.py and data_preprocessing_titanic.py - contain functions used in the script; they are also present in the Jupyter notebook,

Plots:

  • training_iter_*.png - decision boundary during first few steps in training on simulated data,
  • models_comparison_*.png - plots with accuracies of models trained on real datasets.

Datasets used for experiments

  1. Breast Cancer Wisconsin (Diagnostic) Data Set
  2. Titanic: Machine Learning from Disaster