基于Hadoop云的数据库营销海量数据处理与挖掘的研究

发布时间：2018-04-14 07:10

本文选题：数据库营销 + 云计算　；参考：《浙江理工大学》2013年硕士论文

【摘要】：互联网的应用与发展不仅促进了各个新兴产业的产生与发展，影响了每个人的生活，同时也为传统制造业提供了机遇与挑战。对于制造业企业来说，常规的营销方式，如投放广告、搞促销活动等方式已经远远不能满足他们需要。常规营销方式的制定往往需要通过较长的周期及过高的成本收集客户数据然后人工分析，再制定相应的营销策略与方式，甚至在缺少数据的情况下盲目进行营销策略的制定，所以很难达到企业预期的效果。而通过数据库营销，企业可以方便地收集和积累客户信息，构建庞大的顾客信息库，然后通过云计算技术对海量数据快速准确地筛选和分析，从而有效地进行客户数据挖掘与关系维护。本文提出了基于Hadoop云的数据库营销系统的架构，实现海量数据的处理与存储，，并将其应用到红塔集团数据库营销系统中，并且在系统初步完成并投入运行后，红塔集团卷烟销量尤其重点促销品牌销量相比同期的有了大幅度提升，数据库营销在其中起了至关重要的刺激作用。文章主要研究内容如下： 1）分析红塔集团的现状，完成红塔集团数据库营销系统的需求分析，根据需求分析，针对性地对Hadoop分布式计算平台进行研究和综述，了解其优势、架构和运行机制，分析使Hadoop构建红塔集团企业私有云的可行性。 2）探讨了数据库营销常用数据挖掘方法，并根据红塔集团数据库营销系统的实际需求及首要目标，构建了促销活动响应模型，提高集团促销活动客户响应率；构建促销活动决策模型，为决策者提供有效的客户信息，确定促销产品及促销客户群；设计了客户终身价值、客户忠诚度计算方法，构建客户忠诚度预警模型以及客户忠诚度提升模型。 3）研究设计Hadoop与关系型数据库协同工作方案，设计Hadoop分布式平台下MapReduce计算模型对关系型数据的处理方法，并设计使用最优数据集选择算法构建MapReduce Job数据流，实现通用性设计，降低维护成本。 4）根据红塔集团的实际情况，给出了系统总体设计方案，并应用Hadoop构建红塔集团企业私有云。
[Abstract]:The application and development of the Internet not only promote the emergence and development of each new industry, but also provide opportunities and challenges for the traditional manufacturing industry.For manufacturing enterprises, conventional marketing methods, such as advertising, promotional activities and so on, are far from meeting their needs.The formulation of conventional marketing methods often needs to collect customer data through a long period and too high cost, then manually analyze, and then formulate corresponding marketing strategies and methods, even blindly make marketing strategies in the case of lack of data.So it is difficult to achieve the desired results.Through database marketing, enterprises can easily collect and accumulate customer information, build a huge customer information base, and then quickly and accurately screen and analyze massive data through cloud computing technology.In order to effectively carry out customer data mining and relationship maintenance.In this paper, the architecture of database marketing system based on Hadoop cloud is put forward, which realizes the processing and storage of massive data, and applies it to the database marketing system of Hongta Group, and after the system is initially completed and put into operation,Hongta Group's cigarette sales, especially focused on the promotion of brand sales compared with the same period has a significant increase, database marketing has played a vital role in the stimulus.The main contents of this paper are as follows:1) analyzing the current situation of Hongta Group, completing the requirement analysis of Hongta Group's database marketing system, researching and summarizing the Hadoop distributed computing platform according to the requirement analysis, understanding its advantages, structure and operation mechanism.This paper analyzes the feasibility of Hadoop to construct private cloud of Hongta Group.2) the common data mining methods of database marketing are discussed, and according to the actual demand and primary goal of the database marketing system of Hongta Group, the response model of promotional activities is constructed to improve the customer response rate of group promotional activities.The decision model of promotion activities is constructed to provide effective customer information for decision makers, to determine the promotion products and customer groups, and to design a method for calculating customer lifetime value and customer loyalty.Build customer loyalty warning model and customer loyalty promotion model.3) study and design the cooperative work scheme between Hadoop and relational database, design the method of MapReduce computing model to deal with relational data under Hadoop distributed platform, and design the MapReduce Job data stream using the optimal data set selection algorithm to realize the universal design.Reduce maintenance costs.4) according to the actual situation of Hongta Group, the overall design scheme of the system is given, and the private cloud of Hongta Group Enterprise is constructed by using Hadoop.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP311.13

【参考文献】