摘 要
基于大数据的推荐系统的实现主要基于hadoop的mapreduce程序,利用数学上的共线矩阵算法,来求出商品之间的亲密度,这个是要由用户购买的历史数据,经过分析求出来的。现在大数据时代已经到来,现在比较流行的就是hadoop和spark,hadoop是针对于离线数据的分析,而spark可以实时的对数据作出分析,还有一种技术是storm,它也可以做到实时对数据做出分析,但是它具有不稳定性,所以大多数公司都不用它,目前我所知道的只有阿里巴巴在用storm,因为阿里巴巴有一支团队专门研究storm的,他们在storm研究上已经有所突破,但他们对外是保密的,所以别的公司都用spark和hadoop。
我开发的这个系统,采用了C/S结构,前端页面与后台实时连接,同时与数据库也实时连接,我采用的是Mysql数据库,我之所以选择它,是因为Mysql数据库是一个免费的数据库,而且很好用,内存占用也特别小,好多公司都在使用Mysql数据库。在推荐算法部分,我选择的是免费开源的hadoop,因为它配置简单,开源,免费,用Java编程,网上教程多。登录注册部分,我采用的是Spring Mvc框架,因为它只需要简单的配置就可以实现很多功能,如映射SQL语句,JDBC的配置文件,获取Servlet,扫描包文件,很容易就能实现登录和注册功能。页面部分,我采用的是CSS/DIV结合,还有JS设计出动感,优美的页面。
关键词:推荐系统;C/S结构;Mysql数据库;JSP;Spring MVC;Servlet
ABSTRACT
The implementation of the recommendation system based on large data is mainly based on the map algorithm of hadoop, and the mathematic collinearity matrix algorithm is used to find the intimacy between the goods. This is the historical data to be purchased by the users. Now the big data age has come, now more popular is hadoop and spark, hadoop is for offline data analysis, and spark can be real-time analysis of the data, there is a technology is storm, it can also be real-time data But it is not stable, so most companies do not use it, at present I know only Alibaba in the storm, because Alibaba has a team specializing in storm, they have been in the storm study has been Break, but they are confidential, so other companies are using spark and hadoop.
I developed the system, using the C/S structure, front-end pages and background real-time connection, and the database is also real-time connection, I use the Mysql database, I chose it, because the Mysql database is a free database, And very easy to use, memory footprint is also particularly small, many companies are using the Mysql database. In the recommended algorithm section, I chose the free open source hadoop, because it is simple to configure, open source, free, with Java programming, online tutorials. Login part of the registration, I use the Spring Mvc framework, because it only needs a simple configuration can achieve a lot of features, such as mapping SQL statements, JDBC configuration files, access to Servlet, scan package files, it is easy to achieve login and registration Features. Page part, I use the CSS / DIV combination, as well as JS design dynamic, beautiful page.
Keywords: Recommendation System; C/S structure; Mysql database;JSP;Spring MVC; Servlet
目 录
第一章 绪论 1
1.1 研究背景和意义 1
1.2 研究目标 2
1.3 论文结构 3
1.4 本章小结 4
第二章 相关技术与方法说明 5
2.1 架构概述 5
2.2 有关技术简介 5
2.3 开发工具和环境 6
2.4 本章小结 7
第三章 系统分析 8
3.1 系统功能需求分析 8
3.2 用户管理用例图 9
3.3 推荐模块 10
3.4 购物车模块 11
3.5 数据库设计 12
3.6 本章小结 15
第四章 系统详细设计 16
4.1 用户登录和注册模块 16
4.2 系统主页设计 21
4.3 推荐模块的设计 22
4.4 购物车模块的设计 26
4.5 本章小结 26
第五章 系统测试 27
5.1 系统测试综述 27
5.2 测试用例 27
5.3 测试分析 30
5.4 本章小结 30
第六章 总结 31
6.1 创新点 31
6.2 系统优点 31
6.3 系统缺点 31
6.4 存在问题及拟解决方案 31
参考文献 33
致 谢 34