Awesome

Why Model Why - Assessing the Usability of LIME

This repository includes all the files to replicate the work done in the publication "Why model why? Assessing the strengths and limitations of LIME".

Code style

All the work has been conducted using python.

Description

The Juypter Notebook contains of three parts: i) Data processing ii) Application of models iii) Application of LIME

i) The Rain in Australia Dataset is prepared so different algorithms can be applied on it.

ii) A Decision Tree, Random Forest, Logistic Regression and XG Boost are applied and their performance is evaluated using a confusion matrix and ROC Curve.

iii) The xAI Framework LIME is applied to gain insight into how the models processed the data and weighting the features.