【作者】 张良；

【导师】 张雷；

【作者基本信息】 北京邮电大学 ， 计算机科学与技术， 2010， 硕士

【摘要】 数据挖掘是指从大量的数据中发现潜在的、有用的知识的过程。关联规则挖掘是数据挖掘的主要研究内容,它从大量的数据项中寻找隐藏着的联系或相关性。Apriori算法是关联规则挖掘的一个经典算法,但它解决问题的针对性较强,应用面狭窄。因此,本文提出了基于规则的Apriori算法,大大扩展了Apriori算法的应用范围。随着企业业务种类的不断增多,传统的业务规则配置面向技术人员,具有维护难度大、生命周期较短的缺点,已难以适应业务发展的要求。因此,本文提出了面向业务人员的基于SBVR(Semantics of Business Vocabularies and Business Rules,业务词汇和业务规则的语义)的业务规则配置模型。因此,本文重点研究了基于规则的Apriori算法和基于SBVR的业务规则配置技术,具体工作包括以下几个方面：1、研究了Apriori算法挖掘关联规则的步骤。针对传统算法应用面狭窄这一缺陷,本文提出了一种基于规则的Apriori算法——Rule_Apriori。该算法在原有算法的基础上,加入了一层规则,按照配置的规则进行指导和控制Apriori算法的执行。2、分析了传统的业务规则技术的不足,提出了基于SBVR的业务规则配置模型。这是一种面向业务人员、可灵活配置的业务规则配置模型。包括业务规则的定制、基础服务和后台服务三个模块。3、将基于SBVR的业务规则配置模型应用到套餐资费配置上,实现计费功能。验证了基于SBVR的业务规则配置的灵活性,同时为其他应用提供计费功能。4、将Rule_Apriori算法应用于电信套餐制定中和套餐劝转服务中。电信套餐制定人员根据从套餐订购实例中挖掘出的关联规则,找出用户容易在一起定购的产品,从而为制定套餐提供决策支持信息。套餐劝转服务根据用户的消费详单进行套餐匹配,找出比较适合用户的套餐,从而为用户提供建议。更多还原

【Abstract】 Data mining is the process of discovering potential and useful knowledge in large amounts of data. As the main research content of data mining, association rule mining is to find the hidden links or relevance from a large number of data items. Apriori algorithm is the classical algorithm for association rule mining. However, it can only resolve the problem of specific domain. For this narrow application the paper proposed a new rule based Apriori algorithm, which greatly expanded the scope of application.With the growing number of businesses, the traditional configuration of business rules exposes its defects including technical staff oriented configuration style, difficult maintaining, and short life-cycle. The paper presented a business-oriented rules configuration model based on SBVR (Semantics of Business Vocabularies and Business Rules)Therefore, this thesis focuses on the rule based Apriori algorithm and SBVR based business rules technology, mainly including the following aspects:1. Studied the steps of mining association rules. For the narrow application of traditional algorithm, the paper proposed a rule-based Apriori algorithm, named Rule_Apriori. The algorithm added a rule layer based on the original algorithm, which can guide and control the implementation of the Apriori algorithm according to the configuration of the rules.2. Analyzed the shortcomings of traditional business rules technology, and put forward a SBVR based business rules configuration model, which is business oriented and configurable, including three modules:business rule customization, basic services, and background services.3. Applied SBVR-based business rules technology to billing configuration system, and thus verified the flexibility of business rules based on SBVR.4. Applied the Rule_Apriori algorithm to the telecommunication packages and the service of persuading users to switch packages. The telecommunication packages mined association rules from the users order, found the user-friendly packages, so as to provide decision support when customizing packages. The service of persuading users to switch packages found more appropriate packages for users, according to the details records of users’ consumption, and so as to make package recommendations for the customers.更多还原