面向客户个性化产品配置的关联规则挖掘研究

Mining Configuration Rules for Personalized Product Customization

针对个性化产品配置中难以获取隐性配置知识等问题,提出了基于关联规则挖掘方法来获取产品配置规则的方法,从而便于在配置过程中对客户进行个性化的推荐。基于产品配置的历史销售数据,应用基于双层遗传算法来实现了关联规则挖掘算法,并设计了遗传算法的编码表示和算子操作。最后,以平板电脑的客户配置案例为例,举例说明了所提出方法的有效性。实验表明,与经典的Apriori算法相比,所提出的方法能够自适应地获得规则支持度和置信度的阈值,避免了Apriori人为设置阈值所带来的不足之处,从而能够适用于大数据环境下产品的个性化推荐。

With the aim of acquiring tacit configuration rule knowledge in personalized product configuration,the association mining method is used to infer tacit configuration rule knowledge such that personalized recommendation to customers can be made in the configuration process.With the historical sales data of products,a bi-level genetic algorithm with the rule coding representation and corresponding operators is proposed to obtain the association rules.The effectiveness of the presented approach is illustrated by a case study on tablet computers.Experiments demonstrate that,compared with the Apriori algorithm,the proposed method can adaptively obtain the threshold of rule support and confidence,and thus avoid the shortcomings of Apriori method in manually setting a threshold for rule support and confidence.As a result,the tacit rule knowledge in product configuration can be effectively obtained to facilitate the process of product customization and recommendation.