Conversational recommendation: Theoretical model and complexity analysis

Recommender systems help users find items of interest in situations information overload a personalized way, using needs and preferences individual users. In conversational recommendation approaches, the system acquires an interactive, multi-turn dialog. This is usually driven by incrementally asking about their item features or items. A central research goal this context efficiency, evaluated concerning number required interactions until satisfying found. Today, on dialog efficiency almost entirely empirical, aiming to demonstrate, for example, that one strategy selecting questions ask user better than another given application. work complements empirical with theoretical, domain-independent model recommendation. model, designed cover range application scenarios, allows us investigate approaches formal particularly computational complexity devising optimal interaction strategies. An experimental evaluation empirically confirms our findings.