为了在复杂环境中对各种服务推荐策略进行准确评估以选出最合适的服务策略、提高O2O服务推荐质量,运用计算实验方法,从3个方面对O2O服务推荐策略进行了深入研究,包括服务推荐策略的设计(协同过滤推荐策略、考虑情境的协同过滤推荐策略、综合考虑情境和服务状态的推荐策略)、实验系统构建、服务推荐策略的实验分析。实验结果表明,在需求量不同的场景下,综合考虑情境和服务状态的推荐策略表现最优。
To evaluate various service recommendation strategies in complex environments for selecting the most appropriate service operation strategy and to improve the quality of Online to Offline (O2O) service recommendation,by utilizing the computational experiments method, 020 service recommendation strategies were researched from three aspects, which included the service recommendation strategy design (collaborative filtering recommendation strategy, context-based collaborative filtering recommendation strategy and the recommendation strategy considering context and service status), the construction of experiment system and the experiment evaluation analysis of service operation strategy. Through the computational experiments, the performance of recommendation strategy considering context and service status was optimal under different demand scenarios.