Getting Started

Overview

The purpose of FREx is to provide a structure in which you can develop a recommender systems, especially those that use RDF data and use rule-driven steps to determine what to recommend. FREx uses a pipeline system to process candidate items to recommend, and it also attaches explanations about each step that a candidate went through in the recommendation pipeline. Therefore, outputs of a recommendation pipeline implemented using FREx will have explanations about the processes it went through as well as the context in which the candidates were produced attached as a key piece of the final output.

FREx is NOT a framework that takes in your data and spits out a ranking of items to recommend. Rather, it is a framework in which you would incorporate other recommender algorithms to score or generate candidates. FREx also is NOT a tool that automatically generates explanations. The explanations are created by the developer - FREx helps to provide a structure in which such explanations can be passed along and incorporated through the pipeline.

Installation

To use FREx, you should be using Python version 3.8 or higher. This supports the use of dataclass decorators as well as some quality-of-life improvements surrounding type hints.

You can use FREx by installing it via pip

$ pip install git+https://github.com/solashirai/FREx@master#egg=frex

You will now be able to use FREx in your project like any other Python package, with import frex.