Context Adaptation for Smart Recommender Systems

Contextual factors are considered an important mediator for improving recommender system performance. In the e-commerce sector, such contextual factors as users´ real-time state of mind and budget play a critical role in consumers´ decision-making processes. Using a context-aware approach, the authors´ recommendation model can identify users´ state of mind and budget based on clickstream data. They´ve deployed their model on a French e-commerce website for a comparative A/B test. Results show that usage of the context-aware system is significantly higher than that for the benchmarking system.