Changes in version 1.1.0 (2026-10-30)
CRAN release: 2026-10-05
- Made
irlbaandrecosystemoptional dependencies. SVD and LIBMF now give an installation hint when their required package is missing. - Fixed mailto in vignette.
- Added getting-started guide.
- Fixed spelling and grammar in package.
- Switched to testthat edition 3 and added tests for rating matrices, recommendations, prediction accuracy, and evaluation.
Changes in version 1.0.7 (05/29/2025)
CRAN release: 2025-05-31
- slightly better handling of 0 vs. NA in sparse matrices.
- coercion data.frame -> realRatingMatrix: added drop = TRUE so it works with tidyverse tibbles (reported by youngroklee-ml)
Changes in version 1.0.6 (09/19/2023)
CRAN release: 2023-09-20
- fixed bug in row/colSums call for Matrix 1.6-2 (reported by Mikael Jagan).
- updated deprecated coercion for Matrix
Changes in version 1.0.3 (01/20/2023)
CRAN release: 2023-01-21
Changes in version 1.0.0 (05/27/2022)
CRAN release: 2022-05-27
Bugfixes
- calcPredictionAccuracy now works with negative values for given (all-but-x). A negative value produces an error with instructions.
- We require now proxy version >= 0.4-26 which fixed a conversion bug for cosine similarity.
- RECOM_AR now respects already know items (code provided by gregreich).
- evaluate: keepModel = TRUE now works (bug reported by gregreich).
- Recom_SVD: fixed issue with missing values set to zero (bug reported by jpbrooks@vcu.edu)
Changes
- Ratings of zero are now fully supported. We use .Machine$double.xmin to represent 0 in sparse matrices. zapsmall() can be used to change them back to 0.
- topNList has now a method c() to combine multiple lists.
- RECOM_AR: Ratings are now equal to quality measure used for ranking.
- HYBRIDRECOMMENDER: add “max” and “min” aggregation.
- removeKnownRatings is now sparse.
- RECOM_RANDOM now has parameter range to specify the rating range.
Changes in version 0.2-7 (04/26/2021)
CRAN release: 2021-02-26
Changes in version 0.2-6 (06/16/2020)
CRAN release: 2020-06-17
New Features
- ratingMatrix gained method hasRatings.
- Recommender gained method “HYBRID” to create hybrid recommenders. Now hybrid recommenders can also be used in evaluate().
- similarity gained parameters min_matching and min_predictive.
Bugfixes
- predict for Recommender RANDOM now uses the correct user ids in the prediction (reported by aliko-str).
- fixed weight bug in Recommender UBCF (reported by aliko-str).
- Recommender UBCF now removes self-matches if item ids are specified in newdata. Specifying data in predict is no longer necessary. (reported by aliko-str).
- HybridRecommender now handles NAs in predictions correctly (was handled as 0).
Changes in version 0.2-5 (08/27/2019)
CRAN release: 2019-08-27
Changes
- predict with type “ratingMatrix” now returns predictions for the known ratings instead of replacing them with the known values.
- Recommender methods Popular, AR and RERECOMMENDER now also return ratings for binary data (and thus can be used for HybridRecommender).
- Added a LIBMF-based recommender.
Changes in version 0.2-3 (06/19/2018)
CRAN release: 2018-06-19
Bugfixes
- Fixed bug in ALS_implicit (reported by equalise).
- getData for binaryRatingMatrix data with type “known” and “unknown” preserves now user ids/rownames (reported by Kasia Kulma).
- predict for HybridRecommender now retains user IDs (reported by homodigitus).
- Removed warning about using drop in subsetting ratingMatrices (reported by donnydongchen).
Changes in version 0.2-1 (09/15/2016)
CRAN release: 2016-09-17
New Features
- Added recommender method ALS and ALS_implicit based on latent factors and alternating least squares (contributed by Bregt Verreet).
- Changes in recommendation method AR: Default for maxlen is now 3 to find more specific rules. Parameters measure and decreasing for sorting the rule base are now called sort_measure and sort_decreasing. New parameter apriori_control can be used to pass a control list to apriori in arules.
- The registry now has a reference field.
Changes in version 0.2-0 (05/31/2016)
CRAN release: 2016-06-01
- Added recommender RERECOMMEND to recommend highly rated items again (e.g., movies to watch again).
- Added a hybrid recommender (HybridRecommender).
- realRatingMatrix supports now subset assignment with [.
- RECOM_POPULAR now shows the parameters in the registry.
- RECOM_RANDOM produced now random ratings from the estimated distribution of the available recommendations (from a normal distribution with the user’s means and standard deviation).
- predict now checks if newdata (number of items) is compatible with the model.
- getTopNLists and bestN gained a randomized argument to increase prediction diversity.
- Added getRatings method for topNList.
Changes in version 0.1-9 (05/18/2016)
CRAN release: 2016-05-19
- FIX: rownames of newdata are now preserved in prediction output.
- We use testthat now.
- Normalization now can be done on rows and columns at the same time.
- SVD with column-mean imputation now folds in new users.
- Added Funk SVD (funkSVD and recommender SVDF).
- Added function error measures: MAE, MSE, RMSE, frobenius (norm).
- Jester5k contains now the jokes.
- MovieLense contains now movie meta information.
- topNLists now also contains ratings.
- Removed obsolete PCA-based recommender.
Changes in version 0.1-8 (12/17/2015)
CRAN release: 2015-12-18
- Fixed several problems in the vignette.
- predict for realRatingMatrix accepts now type = “ratingMatrix” to returns a completed rating matrix.
- Negative values for given in evaluationScheme implement all-but-given evaluation.
- Method “SVD” used now EM-based approximation from package bcv.
Changes in version 0.1-7 (7/23/2015)
CRAN release: 2015-07-24
- NAMESPACE now imports non standard R packages.