User-Based KNN
Recommend items liked by users most similar to the target user
User-Based KNN finds the K most similar users to the target user based on their interaction history, then recommends items those similar users have engaged with but the target user has not yet seen.
When to use:
- Recommendation for users with sufficient interaction history
- "Users like you also liked…" style personalization
- Small-to-medium user bases where user-user similarity is computationally feasible
Input: User-item interaction history Output: Ranked list of recommended items based on similar-user preferences
Model Settings (set during training, used at inference)
K (N Neighbors) (default: 40) Number of similar users to aggregate recommendations from.
Similarity Metric (default: cosine) Metric for user-user similarity.
Min Support (default: 1) Minimum shared items required between users to compute similarity.
Inference Settings
No dedicated inference-time settings. Recommendations aggregate items from K most similar users.