Alpenglowhttp://mloss.orgUpdates and additions to AlpenglowenMon, 26 Apr 2021 12:57:51 -0000Alpenglow 1.0.6<html><p>Alpenglow is an open-source recommender systems research framework, aimed at providing tools for rapid prototyping and evaluation of algorithms for time-aware and streaming recommendation tasks. It supports modeling non-stationary environments using prequential evaluation and incremental updating of models. The framework is implemented in C++, and also provides an easy-to-use python API.</p> <p>Features:</p> <ul> <li>various tools for evaluation</li> <li>preconfigured experiments</li> <li>option for embedding traditional periodic retraining in the prequential framework</li> </ul> <p>Implemented models:</p> <ul> <li>matrix factorization (SGD, ALS, iALS)</li> <li>asymmetric matrix factorization</li> <li>SVD++</li> <li>factorization machines</li> <li>nearest neighbor</li> <li>time-aware popularity</li> <li>transition probability</li> </ul> </html>Domokos Kelen, Erzsebet Frigo, Robert Palovics, Levente Kocsis, Andras A. BenczurMon, 26 Apr 2021 12:57:51 -0000 systemonline leaning