Impact of Product Consumption Patterns on Recommender Systems Performance: An Agent-Based Modeling Approach

We develop an agent-based modeling and simulation approach to study the impact of product consumption strategies on the temporal dynamics of recommender systems’ performance. We model the product consumption strategy by decomposing it into two parts: recommender system’s item ranking strategy and consumers’ item selection strategy. Our simulation results show that consumers’ selection strategy has great impact on system’s recommendation quality in subsequent time periods. Interestingly, the more heavily users rely on the system’s recommendations to make item choices, the more inaccurate (and, therefore, arguably less useful) system’s predictions become in the future, which has important implications for recommender systems design.