Large Language Models are Competitive Near Cold-start Recommenders for Language- and Item-based Preferences.
Scott SannerKrisztian BalogFilip RadlinskiBen WedinLucas DixonPublished in: RecSys (2023)
Keyphrases
- language model
- cold start
- recommender systems
- user preferences
- collaborative filtering
- cold start problem
- language modeling
- data sparsity
- information retrieval
- document retrieval
- n gram
- probabilistic model
- user ratings
- matrix factorization
- query expansion
- retrieval model
- test collection
- user profiles
- personalized recommendation
- tag recommendation
- vector space model
- implicit feedback
- natural language
- query terms
- recommendation systems
- recommendation algorithms
- smoothing methods
- linguistic knowledge
- information overload
- text categorization
- multi modal
- feature selection