Eigenvalue analogy for confidence estimation in item-based recommender systems.
Maurizio Ferrari DacremaPaolo CremonesiPublished in: CoRR (2018)
Keyphrases
- recommender systems
- collaborative filtering
- matrix factorization
- information filtering
- item based collaborative filtering
- user preferences
- parameter estimation
- trust aware
- high confidence
- least squares
- information overload
- cold start problem
- data sparsity
- density estimation
- user modeling
- personalized recommendation
- user profiling
- robust estimation
- statistical mechanics
- confidence values
- recommendation algorithms
- estimation accuracy
- data sets
- estimation algorithm
- user interests
- neural network