recommenders / rival

RiVal recommender system evaluation toolkit

Home Page:rival.recommenders.net

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

RiVal

Build Status Join the chat at https://gitter.im/recommenders/rival codecov License

RiVal is a toolkit for data splitting and evaluation of recommender systems. This page contains information on how to work with the RiVal code and how to compile RiVal. For more information about RiVal and the documentation, visit the RiVal website or wiki. If you have not used RiVal before, do check out the Getting Started guide.

RiVal is made available under Apache License, Version 2.0.

Installation and Dependency Management

RiVal is built and deployed via Maven. In order to install it, check out this repository and run mvn install. This will make it available for other Maven projects as a dependency.

Modules

RiVal is comprised of one top-level module and four sub-modules. The top-level rival module is a container module used to build the submodules and provide all needed settings and dependencies. The five sub-modules are as follows:

  • rival-core -- the common data structures and similar object used throughout RiVal.
  • rival-evaluate -- the evaluation module, contains metrics and strategies used for evaluation.
  • rival-examples -- a module containing examples on how to use RiVal programmatically.
  • rival-recommend -- the recommendation modeule, contains hooks to Apache Mahout and LensKit.
  • rival-split -- the data splitting module, contains different data splitting strategies.
  • rival-package -- a configuration module for bulding rival distributions.

Contributing to RiVal

We're very happy if you want to contribute to RiVal. If you are not sure on what to work on, have a look a the current issues.

If you want to contribute, do it in the form of GitHub pull requests. To do this:

  1. Find the issue you want to fix, or create a new one describing what your contribution fixes.
  2. Fork RiVal rival on GitHub
  3. Push your changes to your fork
  4. Submit a pull request via the GitHub Web interface

Note that all your contributions will be licensed under RiVal's copyright licences (Apache). If you are committing a new class, make sure to include the copyright statement at the top.

Mailing list

To subscribe to the rival mailing list, visit and join the rival-users Google group.

More information

About

RiVal recommender system evaluation toolkit

rival.recommenders.net

License:Apache License 2.0


Languages

Language:Java 98.6%Language:Shell 1.4%