A three-class ranking loss that partitions unlabeled data into neutral and negative and improves top-K recommendation across multiple collaborative filtering models.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Unlocking the Hidden Treasures: Enhancing Recommendations with Unlabeled Data
A three-class ranking loss that partitions unlabeled data into neutral and negative and improves top-K recommendation across multiple collaborative filtering models.