Difference between revisions of "LensKit"
(Add to software category) |
Zeno Gantner (talk | contribs) |
||
| Line 1: | Line 1: | ||
| − | + | '''LensKit''' is a Java-based recommender toolkit from [[GroupLens]]. It provides a common API for recommender algorithms, an evaluation framework for offline evaluation of recommender performance, and highly modular implementations of common collaborative filtering algorithms. | |
It is intended to be particularly useful for research and education. The aim of the project is to produce a common recommender framework, reusable in applications, and clear, readable implementations of common algorithms employing best practices with regards to implementation strategies, tuning, and normalizations. | It is intended to be particularly useful for research and education. The aim of the project is to produce a common recommender framework, reusable in applications, and clear, readable implementations of common algorithms employing best practices with regards to implementation strategies, tuning, and normalizations. | ||
LensKit is under active development. The project currently has several rating-based algorithms and is working on developing additional algorithms, basket/history-based algorithms, and additional evaluation strategies. | LensKit is under active development. The project currently has several rating-based algorithms and is working on developing additional algorithms, basket/history-based algorithms, and additional evaluation strategies. | ||
| + | |||
| + | == External links == | ||
| + | * http://lenskit.grouplens.org | ||
[[Category:Software]] | [[Category:Software]] | ||
Revision as of 06:03, 17 June 2011
LensKit is a Java-based recommender toolkit from GroupLens. It provides a common API for recommender algorithms, an evaluation framework for offline evaluation of recommender performance, and highly modular implementations of common collaborative filtering algorithms.
It is intended to be particularly useful for research and education. The aim of the project is to produce a common recommender framework, reusable in applications, and clear, readable implementations of common algorithms employing best practices with regards to implementation strategies, tuning, and normalizations.
LensKit is under active development. The project currently has several rating-based algorithms and is working on developing additional algorithms, basket/history-based algorithms, and additional evaluation strategies.
