caionobrega / explaining-recommendations

Explaining Recommendations Through Local Surrogate Models

Explaining Recommendations Through Local Surrogate Models

This repository contains the code to run the experiments present in this paper

Setup Environment

1. Install Anaconda Distribution for Python 3.7;
2. Create a virtual environment: conda env create -f env.yml;
3. Activate the virtual environment: conda activate sac2019_env;

NOTE:

• deactivate the virtual environment run: conda deactivate sac2019_env.

Dataset

The folder dataset should have the following files (with respective headers):

• training: <user_id>\t<item_id>\t
• test: <user_id>\t<item_id>\t
• item_features: <item_id>\t<feature_name>\t

The dataset folder path should be set in experiment/utils.py

Folder Structure

• src folder:
  

  • classes and utility functions to be able to run the experiments.

• experiment folder:
  

  • the setup subfolder has the code to generate the explanations;
  • the other scripts computes metrics