#15: Podcast Recommendations in the ARD Audiothek with Mirza Klimenta
Recsperts - Recommender Systems Experts
00:00
Exploring Hybrid Recommendation Systems
Delving into the fusion of content-based and collaborative filtering for hybrid recommendation systems, this chapter covers the challenges of personalized recommendations, signal optimization, and user preference integration. The conversation further extends to models like matrix factorization, hyperpolarity bias, and graph-based approaches for enhanced accuracy in recommending podcast programs. Additionally, the chapter touches on optimizing recommenders for diversity, popularity, and fairness, including the use of inverse propensity weighting and novel methods to reduce bias.
Transcript
Play full episode
