基于Agent的顾客行为及个性化推荐仿真模型

针对传统个性化推荐研究方法的局限性, 提出一种基于Agent建模与仿真的方法, 通过个体的交互作用所产生的涌现特征来分析移动电子商务环境下的顾客行为及个性化推荐策略的有效性． 以移动商务环境下的餐饮推荐系统为例, 分析了消费活动过程中顾客与服务的交互行为, 以及情境因素对顾客消费的影响, 构建了服务推荐及顾客行为规则, 并在REPAST环境下实现了本Agent仿真模型． 仿真结果表明:该模型可有效分析及预测服务推荐和顾客决策的涌现现象, 并由此推断顾客总体的消费趋势; 同时, 考虑情境因素的推荐模型的有效性比单独基于顾客个性化信息的推荐模型有明显提高．

Agent-based simulation model of customer behavior and personalized recommendation

This paper proposes an agent-based modeling and simulation method to overcome the limitations of traditional personalized recommendation method. Customer behavior and the effectiveness of personalized recommendation strategy under mobile electronic commerce were analyzed according to the emergence generated by the interactions of each agent entity. Taking catering recommendation system under mobile e-commerce as an example, this paper analyzed the interactions of customer and server in consuming process and the impact of context to customer consuming, and the recommendation and customer behavior rules were built, then on this basis the agent simulation model was realized under REPAST simulation environment. The simulation results show that this model can analyze and forecast the emergence of server recommendation and customer decision, and accordingly deduce the general consumer trends. Moreover, the effectiveness of recommendation model considering context is enhanced obviously compared with that of model based on only customer personalization information.