The document summarizes a doctoral thesis on context-aware recommender systems for opportunistic environments. Key points:
1) The thesis proposes a novel context-aware recommender system (CARS) solution designed specifically for opportunistic environments where devices can communicate directly.
2) A tag-based approach is presented that uses user-defined tags to characterize both user context and item information, allowing a single multi-domain recommender system to be built.
3) An algorithm called PLIERS is introduced that recommends items based on tag popularity in a way that does not require parameter tuning and increases neither computational complexity nor recommendation bias.