用户建模是电子商务领域的重要研究问题之一,但是到目前为止还没有非常有效的方法来对用户的动态偏好等特征进行建模.文中在分析电子商务领域用户偏好特征获取的一般特点的基础上,提出基于实例的两阶段递增式用户偏好特征获取方法.同时采用基于半环的软约束满足问题来表示用户的各类偏好特征,将用户建模看成是在潜在方案空间上施加一系列偏好约束的过程.文中还提出了不同的偏好约束类型以及利用偏好约束建立用户模型的过程,实现了电子商务领域复杂用户偏好的灵活建模.
Although no effective methods preference elicitation the user modeling is one of the important research topics in electronic commerce, there are still to model the dynamic preferences of users. In this paper, the general characteristics of user are analyzed and a two-stage case-based incremental preference elicitation approach is proposed. Semiring-based soft constraints are then used to represent various user preferences by considering the user modeling as an imposed process of preference constraints in decision spaces. Moreover, different types of preference constraints and the process of user preference modeling are proposed. By the above-mentioned statements, the flexibly modeling of dynamic preferences in electronic commerce is finally implemented.