Context Adaptation for Smart Recommender Systems.

International audienceContextual factors are considered as important mediator to improve the performance of a recommender system. In e-commerce sector, such contextual factors as users’ real time state of mind and budget play a critical role in the consumer decision-making process. Using a context-aware approach, we develop a recommendation model which can identify user’s state of mind and budget based on clickstream data. The model is deployed on a French e-commerce website for comparative A/B test. Result shows that usage of the context-aware system is significantly higher than the benchmarking system. Implications and managerial suggestions are discussed