gszep/recommendation.jl

Building recommender systems in Julia

Recommendation.jl

Recommendation.jl is a minimal, customizable Julia package for building recommender systems. Pre-built basic functionalities include:

• Non-personalized baselines that give unsophisticated, rule-based recommendation.
• Collaborative filtering on either explicit or implicit user-item matrix.
• Model-based factorization approaches such as Singular Value Decomposition (SVD), Matrix Factorization (MF), and Factorization Machines (FMs).
• Content-based filtering by using the TF-IDF weighting technique.
• Evaluation based on a variety of rating and ranking metrics, with easy-to-use N-fold cross validation executor.

Installation

julia> using Pkg; Pkg.add("Recommendation")

Usage

This package contains a unified DataAccessor module and several non-personalized/personalized recommenders, as well as evaluation metrics such as Recall:

See Getting Started in documentation for the details.

Development

Change the code and test locally:

julia> using Pkg; Pkg.activate(@__DIR__); Pkg.instantiate()
julia> Pkg.test("Recommendation")

Note that unit tests for dataset loaders (e.g., load_movielens_latest()) are conditionally triggered as follows, so that CI does not make excessive download requests to the external sites:

julia> Pkg.test("Recommendation", test_args=["data", "download"])

Build documentation contents:

$ julia --project=docs -e 'using Pkg; Pkg.develop(PackageSpec(path=pwd())); Pkg.instantiate()'
$ julia --project=docs docs/make.jl
$ open docs/build/index.html

Follow JuliaRegistries/Registrator.jl for releasing.