Interpretable_models

General library for implementing ML models and performing numerous interpretation of model features, data attributes, input and output data.

By : Vishwas Sathish
E-Mail : vishwassathish@gmail.com

Requirements :

• python3.x (tested on python3.6)
• sklearn (tested on version 0.19.1)
• pandas (tested on version 0.23.1)
• matplotlib (tested on version 2.2.2)
• lime
• tkinter (gui) / jupyter notebook

--- Get into interpretable_models folder ---

• "codes" folder contains raw python source code for all logistic regression analysis. running the file "log_regression_final.py" as "$> python log_regression_final.py" on command line will render a GUI and instructions can be easily followed from there.

• jupyter notebook files contains the notebook on which we have done our analysis. If jupyter notebook is available in your system, upload "log_regression.ipynb" and run it as a notebook.

• This file also has the markdown and html format of the notebook, which can only be used to view the code snippets and their corresponding outputs.

• "graphs" contains the final graphs for each kind of analysis

  1. Weight analysis
  2. Weight * Value analysis
  3. Lime analysis

• READ THE TEXT FILE "./codes/Instructions_to_run_files" to find out how to run files and obtain explanations/graphs.