摘要
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.
This paper proposes an agent-based modeling and simulation method to overcome the limi- tations 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 cus- tomer 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 emer- gence 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.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2013年第2期463-472,共10页
Systems Engineering-Theory & Practice
基金
国家自然科学基金重大项目(70890080
70890083)
关键词
个性化推荐
Agent建模与仿真
情境
顾客行为模型
有效性评估
personalized recommendation
agent-based modeling and simulation
context
customer be-havioral model
effectiveness evaluation
作者简介
金淳（1963-），男，教授，博士生导师，研究方向：物流与供应链管理、系统仿真、商务智能；
张一平（1989-），女，硕士研究生，研究方向：系统仿真、商务智能．
