RecPack: An(other) Experimentation Toolkit for Top-N Recommendation using Implicit Feedback Data
Lien Michiels, Robin Verachtert, Bart Goethals
Abstract
RecPack is an easy-to-use, flexible and extensible toolkit for top-N recommendation with implicit feedback data. Its goal is to support researchers with the development of their recommendation algorithms, from similarity-based to deep learning algorithms, and allow for correct, reproducible and reusable experimentation. In this demo, we give an overview of the package and show how researchers can use it to their advantage when developing recommendation algorithms.
Topics & Concepts
Computer scienceRecommender systemSimilarity (geometry)ExtensibilityInformation retrievalData miningHuman–computer interactionMachine learningArtificial intelligenceProgramming languageImage (mathematics)Recommender Systems and TechniquesAdvanced Bandit Algorithms ResearchMachine Learning in Healthcare