【作者】 张小刚；

【导师】 刘志镜；

【作者基本信息】 西安电子科技大学 ， 计算机应用技术， 2009， 硕士

【摘要】 OLAP是共享多维信息的、针对特定问题的联机数据访问和分析的快速软件技术。本文首先研究了数据仓库的相关知识与技术,分析处理和传统的事务处理有着很多不同的地方,直接使用数据库技术来进行信息分析是不合适的,需要专门的针对信息分析的技术,数据仓库的概念的产生正是基于这个目的。接着主要研究了OLAP的相关技术,其中介绍了OLAP的两种主要类型：MOLAP和ROLAP及对提升OLAP性能方面一些比较重要的技术。然后讨论了数据立方体中的重要概念视图的相关知识,并对视图的物化选择中现有算法的问题,提出了通过分析分析人员兴趣来对视图物化的选择提供帮助,在基于强化学习的思路上讨论了如何进行兴趣发现,最后针对其中的主要算法提出一种新的思路进行改进。综合前面讨论的各项知识和技术,根据实验室课题设计了一个面向社区服务的数据分析系统,并应用在利安公司的商业智能系统中。最后对本系统做了相关的总结和展望。更多还原

【Abstract】 On-line Analysis Processing (OLAP) is an approach to answer multi-dimensional analytical queries quickly. OLAP is a part of the broader category of business intelligence, which also encompasses relational reporting and data mining. This paper researches on the knowledge and technologies of data warehouse firstly, which is used for the data analysis where common database is improper due to the differences between the analysis processing and the transaction processing. Then the technologies about OLAP are studied, including the character of the two main types of OLAP: MOLAP and ROLAP. Consequently, the view in the data cube is discussed, and a new thought using user interest for the selection of materialized views is studied. How to find the user interest is introduced in the way of reinforcement learning (RL). An innovative algorithm is proposed for Q-learning which is the highly weighted in the architecture of the RL.A Community Service Data Analysis System is designed and implemented in the business intelligence system of Line Corporation with the knowledge and technologies mentioned above. The last part of this paper is the conclusion and prospect about this system.更多还原