Covers: theory of Hybrid Recommendation Systems
Estimated time needed to finish: 5 minutes
Questions this item addresses:
  • What are hybrid Systems?
  • Why do we need hybrid systems?
How to use this item?

Read Section 6.1 It would be useful if you are able to understand the following

  1. What is cold-start problem ( and why knowledge-based systems outperform content-based or collaborative systems?
  2. What kind of models perform better with various kinds of data that are available.
  3. Have a basic understanding of bias and variance of a machine learning system. In short bias represents how restrictive a particular model is whereas variance represents how finicky the model is to random variations of the data. ( - explains this in detail)
Author(s) / creator(s) / reference(s)
Charu Aggarwal, internet
0 comment

Ensemble Based And Hybrid Recommender Systems

Total time needed: ~5 minutes
Gives a basic idea of combining multiple individual recommendation systems to build a robust high-performing Recommendation System
Potential Use Cases
Build a Hybrid Recommendation System that combines the strengths and avoids the weakness of different techniques for building a Recommendation System.
Who is This For ?
INTERMEDIATEPeople with familiarity with various Recommendation System Building techniques like collaborative filtering, knowledge and content based recommendations,
Click on each of the following annotated items to see details.
ARTICLE 1. Hybrid Recommendation Systems
  • What are hybrid Systems?
  • Why do we need hybrid systems?
5 minutes
ARTICLE 2. Ensemble Based Recommendation Systems
  • What are ensemble based systems and why do we need them?
  • What is a Weighted Ensemble?
  • What is a Switching Ensemble?
  • What is a Cascade Ensemble?
  • What is a Feature Augmentation Ensemble?
10 minutes
ARTICLE 3. Monolith Recommendation Systems
  • What are monolith based Recommendation Systems?
  • What is a Meta-level Hybrid Recommendation System?
  • What is Feature Combination Hybrid Recommendation System?
10 minutes
ARTICLE 4. Mixed Recommendation Systems
  • What are mixed Hybrids? Why do we need them?
10 minutes

Concepts Covered

0 comment