Cornac

New features and improvements

- Add `FilmTrust` dataset (266)
- Update `MCF`, `C2PF`, and `SoRec` models for the compatibility (261, 262)
- Support retrieving node degree in `GraphModality` (267)

New features and improvements

- Support `UIRT` data format (252)
- Fix bug in AUC metric (256)
- Tutorial on contributing an evaluation metric (258)

New models

- Multi-Task Explainable Recommendation (MTER) (238)
- Most Popular (MostPop) ranking baseline (244)
- Global Average (GlobalAvg) rating baseline (245)

New features and improvements

- Update VAECF model (237)
- `normalize()` function in utils (248)
- `TfidfVectorizer` in `TextModality` (249)

New features and improvements

- Improve documentation of models (239)
- Update BPR model with removing scaling in `score()` function (240)

New features and improvements

- Fix bug of list multiplication in `ranking_eval()` function (236)

New models

- Explicit Factor Model (EFM) (181)

New features and improvements

- Rename model CF to WMF (226)
- Remove usage of old Dataset from IBPR model (227)
- Use Cornac's mini-batch iterator in PCRL (229)
- Support early stopping for GMF, MLP, and NeuMF (233)

- Support item negative sampling based on popularity (228)
- Remove data_utils.py (230)
- Add Amazon Toys and Games dataset (179)
- Add Sentiment Modality (180)
- Add binary argument into uir_iter() (231
- Move rating and ranking eval functions out for reuse (232)