ModelAssistant-Replication-Package

Overview

This repository contains the replication package for the article entitled "Using graph-based structures in intelligent modeling assistants: an experience report"

Features

• The repo is structured as follows.
• The folder MORGAN contains the source code of the tool used in the paper
• For the BORA tool please refer to the corresponding repository available here
• The dataset used in the evaluation are stored in the Dataset.zip folder

Installation

To run MORGAN please follows the following steps:

1. Clone the repository:

   git clone https://github.com/claudioDsi/ModelAssistant-Replication-Package.git

2. Navigate to the project directory:

   cd ModelAssistant-Replication-Package

3. Install the dependencies from the requirement.txt file:

   pip install -r /path/to/requirements.txt

Please note that Python 3.7 is required for the Grakel library.

Usage

To run MORGAN, you need to run the following steps:

python main.py data_path n_classes n_items size rec_type

where:

• data_path: (string) Path to the dataset folder containing the train and test files.
• n_classes: (integer) Number of classes for recommendation.
• n_items: (integer) Number of items to process for each recommendation.
• size: (integer) Size of the test according to different configurations
• rec_type: (string) Type of recommendation (class or attrs)

To compute the similarity metrics, you can use the compute_similarity function in the main.py by specifying the source data, i.e., one of the three dataset contained in teh zip file, and the output CSV name.