Improving code recommender systems by incorporating domain knowledge and graphical models.
Marcel BruchMira MeziniPublished in: RSSE@SIGSOFT FSE (2008)
Keyphrases
- graphical models
- domain knowledge
- recommender systems
- belief propagation
- probabilistic model
- random variables
- probabilistic inference
- probabilistic graphical models
- approximate inference
- bayesian networks
- collaborative filtering
- conditional random fields
- structure learning
- belief networks
- exact inference
- conditional independence
- map inference
- markov networks
- prior knowledge
- loopy belief propagation
- factor graphs
- background knowledge
- graph structure
- training data
- pairwise
- structural learning