Recommender Systems Algorithms §

Recommender System Algorithms#moc#algorithm

Overview §

Algorithms §

• Content based Filtering
• BPRMF (Bayesian Personalized Ranking with Matrix Factorization, for implicit Feedback)
• ItemKNN (Item-based top-N recommendation algorithm)
• Item Attribute KNN
• UserKNN (User Attribute top-N)
• User Attribute KNN
• SVD (Singular Value Decomposition)
• SVD++
• gSVD++
• FISM: Factored Item Similarity Models for Top-N Recommender Systems

See https://recbole.io/docs/user_guide/model_intro.html#general-recommendation

---

Collaborative Filtering Systems §

Memory Based CF §

User Based CF Algorithms §

• Nearest Neighbour … find items of similar users (Scales bad)

Item Based CF Algorithms §

• Nearest Neighbour … find similarly rates items from other users (scales well)

Model Based CF Algorithms §

• Slope One
• Matrix Factorization … User/Item Matrice
• Deep Learning
• Association Rules
• Probabilistic Models
• Decision Tree (DT)
  • Random Forest (RF) … Train each tree independently, using a random sample of the data.
  • Gradient Boosting (CBT)
  • XGBoost … trains multiple forests