| Имя | Описание |
|---|---|
| AlbumCounter | Predict item that has the album that a user accessed most |
| ArtistCounter | Predict item that has the artist that a user accessed most |
| AttributeCounter | Simple recommender that counts attributes |
| BPRMF | Matrix factorization model for item prediction (ranking) optimized for BPR |
| BPRMF_KDD | BPRMF with frequency-adjusted sampling, prototype for KDD Cup 2011 |
| BPR_Linear | Linear model optimized for BPR |
| BPR_SMF_KDD | BPRMF with frequency-adjusted sampling and shared factors |
| BalancedLogisticRegressionMatrixFactorization | Matrix factorization model optimized for balanced logistic regression. |
| FilterBPRMF | |
| GenreCounter | Predict item that has the genres that a user accessed most |
| ItemAttributeSVM | Content-based filtering using one support-vector machine (SVM) per user |
| ItemKNN | Unweighted k-nearest neighbor item-based collaborative filtering using cosine similarity |
| ItemRecommender | Abstract item recommender class that loads the training data into memory |
| KNN | Base class for item recommenders that use some kind of k-nearest neighbors (kNN) model |
| MF | Abstract class for matrix factorization based item predictors |
| MostPopular | Most-popular item recommender |
| Prediction | Class that contains static methods for item prediction |
| SoftMarginRankingMF | Matrix factorization model for item prediction optimized for a soft margin (hinge) ranking loss, using stochastic gradient descent (as in BPR-MF). |
| TransductiveBPRMF | BPR variant that takes into account some test data for training |
| TransductiveBPRMF_KDD | Transductive BPRMF with frequency-adjusted sampling, prototype for KDD Cup 2011 |
| TypedBPRMF | BPRMF with frequency-adjusted sampling and shared factors |
| UserKNN | k-nearest neighbor user-based collaborative filtering |
| WRMF | Weighted matrix factorization method proposed by Hu et al. and Pan et al. |
| WRMF_KDD | Weighted matrix factorization method proposed by Hu et al. and Pan et al.; adapted to KDD Cup 2011 |
| WeightedItemKNN | Weighted k-nearest neighbor item-based collaborative filtering using cosine similarity |
| WeightedUserKNN | Weighted k-nearest neighbor user-based collaborative filtering using cosine-similarity |
| Zero | Constant item recommender for use as experimental baseline. Always predicts a score of zero |