Recommender Systems

  1. Collaborative Filtering
    1. Core Principles
      1. Leveraging Collective User Behavior
        1. Assumptions of Similarity
          1. Wisdom of the Crowd
          2. Memory-Based Collaborative Filtering
            1. User-Based Collaborative Filtering
              1. Finding Similar Users
                1. Similarity Metrics
                  1. Pearson Correlation
                    1. Cosine Similarity
                      1. Jaccard Similarity
                        1. Adjusted Cosine Similarity
                        2. Neighborhood Selection Strategies
                          1. Top-K Neighbors
                            1. Threshold-Based Selection
                          2. Predicting Ratings
                            1. Weighted Average of Neighbor Ratings
                              1. Mean-Centered Predictions
                                1. Normalization Techniques
                                2. Generating Recommendations
                                  1. Top-N Recommendation
                                    1. Handling Ties and Ranking
                                  2. Item-Based Collaborative Filtering
                                    1. Finding Similar Items
                                      1. Item-Item Similarity Computation
                                        1. Similarity Matrix Construction
                                        2. Building Item-Item Similarity Matrix
                                          1. Precomputation Strategies
                                            1. Storage Considerations
                                            2. Predicting Ratings
                                              1. Aggregating Ratings from Similar Items
                                                1. Weighted Combination
                                                2. Generating Recommendations
                                                  1. Top-N Recommendation
                                                    1. Score Aggregation
                                                  2. Strengths and Weaknesses of Memory-Based CF
                                                    1. Simplicity and Interpretability
                                                      1. Scalability Issues
                                                        1. Sensitivity to Sparsity
                                                          1. Cold-Start Limitations
                                                        2. Model-Based Collaborative Filtering
                                                          1. Rationale
                                                            1. Overcoming Sparsity Issues
                                                              1. Improving Scalability
                                                                1. Noise Reduction
                                                                2. Matrix Factorization Techniques
                                                                  1. Singular Value Decomposition
                                                                    1. Low-Rank Approximation
                                                                      1. Truncated SVD
                                                                      2. Probabilistic Matrix Factorization
                                                                        1. Bayesian Approaches
                                                                          1. Gaussian Assumptions
                                                                          2. Non-negative Matrix Factorization
                                                                            1. Non-negativity Constraints
                                                                              1. Interpretability Benefits
                                                                              2. Alternating Least Squares
                                                                                1. Optimization Approach
                                                                                  1. Parallel Implementation
                                                                                2. Latent Factor Models
                                                                                  1. User Latent Factors
                                                                                    1. Item Latent Factors
                                                                                      1. Predicting Ratings via Dot Product
                                                                                        1. Regularization Techniques
                                                                                          1. L1 Regularization
                                                                                            1. L2 Regularization
                                                                                          2. Training Model-Based Methods
                                                                                            1. Loss Functions
                                                                                              1. Mean Squared Error
                                                                                                1. Root Mean Squared Error
                                                                                                  1. Regularized Loss Functions
                                                                                                  2. Optimization Algorithms
                                                                                                    1. Stochastic Gradient Descent
                                                                                                      1. Alternating Least Squares
                                                                                                        1. Mini-Batch Optimization
                                                                                                        2. Hyperparameter Tuning
                                                                                                          1. Learning Rate Selection
                                                                                                            1. Regularization Parameter Tuning
                                                                                                            2. Model Evaluation and Validation
                                                                                                              1. Cross-Validation
                                                                                                                1. Holdout Validation

                                                                                                          Previous

                                                                                                          3. Content-Based Filtering

                                                                                                          Go to top

                                                                                                          Next

                                                                                                          5. Hybrid Recommender Systems