Table 2 Hybrid filtering methods
From: A systematic review and research perspective on recommender systems
Hybrid methods | Description |
|---|---|
Meta-level | A pre-learned model is used as an input to another recommender system |
Feature combination | The features of one recommender system are injected into another |
Feature augmentation | The result of one model is applied as an input to another |
Mixed hybridization | The output of different recommender systems are mixed, and the combined result is given as a recommendation |
Cascade hybridization | One system improvises the output of another |
Switching hybridization | Select one recommender model based on the current requirement |
Weighted hybridization | Ratings of different techniques are aggregated to compute a single recommendation |