MyMediaLitehttp://mloss.orgUpdates and additions to MyMediaLiteenTue, 08 Oct 2013 22:29:29 -0000MyMediaLite 3.10<html><p>MyMediaLite is a lightweight, multi-purpose library of recommender system algorithms. </p> <p>It addresses the two most common scenarios in collaborative filtering: </p> <ul> <li> rating prediction (e.g. on a scale of 1 to 5 stars), and </li> <li> item recommendation from positive-only feedback (e.g. from clicks or purchase actions). </li> </ul> <p>MyMediaLite gives you a choice of many recommendation methods: </p> <ul> <li> dozens of different recommenders </li> <li> methods can use collaborative and attribute/content data </li> </ul> <p>MyMediaLite is ready to use: </p> <ul> <li> MyMediaLite includes evaluation routines for rating prediction and item prediction; it can measure MAE, NMAE, RMSE, CBD, AUC, prec@N, MAP, NDCG, MRR. </li> <li> It also comes with command line tools for both recommendation tasks that read a simple text-based files. </li> </ul> <p>MyMediaLite is compact: The core library has a size of about 150KB. </p> <p>Portability: Written in C#, for the .NET platform; runs on every architecture supported by Mono: Linux, Windows, Mac OS X. </p> <p>Freedom: MyMediaLite is free software/open source software. It can be used, modified, and distributed under the terms of the GNU General Public License (GPL). </p> <p>Additional features: </p> <ul> <li> Serialization: save and reload recommender models </li> <li> Real-time incremental updates for many recommenders </li> <li> multi-core support </li> </ul></html>Zeno Gantner, Steffen Rendle, Lucas Drumond, Christoph FreudenthalerTue, 08 Oct 2013 22:29:29 -0000 based learninglarge scale learningalgorithmsdata miningevaluationsupervised learningcollaborative filteringmatrix factorizationrecommender systemsknnlibrarydotnetmono