面向客户群的服务方案配置规则挖掘方法

Customer Group Based Rule Mining Approach for Service Concept Configuration

产品/服务配置规则获取的主要方式是用数据挖掘技术从设计实例数据库中提取.客户化服务方案配置规则为服务功能需求和方案特征间的关联规则.考虑到常用关联规则挖掘算法Apriori具有运算复杂的缺点,提出基于PIETM(Principle of Inclusion—Exclusion and Transaction Mapping)算法的配置规则挖掘方法,考虑置信度和有趣度指标,提取强关联规则.针对配置实例数据库数据量较大时,配置规则挖掘的效率会降低且会产生大量冗余规则的问题,采用二元语义模型表达定性的服务功能需求,将同类客户群的功能需求进行合并,替换多样化的功能需求,减少规则的冗余.最后以一工程机械制造企业服务方案配置规则挖掘为例,验证了所提方法的有效性.

The main approach of acquiring product or service configuration rules is to extract the configuration rules from design knowledge database by using data mining methods. Configuration rules of customization service concept design are the association rules between function requirements and concept characteristics. Considering Apriori which is a common association rule mining algorithm has the drawback of complicated operation, a configuration rules mining method based on PIETM （Principle of Inclusion--Exclusion and Transaction Mapping） is proposed, and the strong association rules were extracted according to the confidence degree and the interestingness degree. The efficiency of mining configuration rules would decrease and a large amount of redundant rules will be obtained when the configura- tion design instances database is large. Aiming at this problem, two - tuple linguistic model was adopted to express qualitative service functional requirements, the functional requirements of the same customer group were combined and multiple functional requirements were replaced, then the quantity of redundant rules was decreased. Finally, a case study of rule mining for service concept configuration design was presented to illustrate the effectiveness of the proposed approach.