定制化母婴用品推荐系统

论文总字数：23663字

目 录

摘要 ………………………………………………………………………1

Abstract …………………………………………………………………2

1 本文绪论 ……………………………………………………………3

1.1 课题研究背景……………………………………………………………………………3

1.2 国内外研究现状…………………………………………………………………………4

1.3 课题研究的目的…………………………………………………………………………4

1.4 本文的结构及流程………………………………………………………………………5

2 推荐系统的用户接口…………………………………………………6

2.1 推荐系统用户接口的发展概述 ………………………………………………………6

2.1.1 电子商务平台……………………………………………………………………6

2.1.2 电子商务的历史及特点…………………………………………………………6

2.1.3 电子商务平台的应用模式………………………………………………………7

2.2 我国电商系统的发展简介………………………………………………………………8

2.2.1 我国电子商务系统发展历史……………………………………………………8

2.2.2 我国电子商务系统发展前景……………………………………………………8

3 推荐系统算法以及架构简介…………………………………………9

3.1 推荐系统算法简介………………………………………………………………………9

3.1.1 协同过滤推荐算法………………………………………………………………9

2.1.2 基于内容的推荐算法 …………………………………………………………11

3.1.3 基于人口统计学的推荐算法 …………………………………………………11

3.1.4 混合推荐算法……… …………………………………………………………12

3.1 推荐系统架构 …………………………………………………………………………12

3.2.1 推荐系统整体结构 ……………………………………………………………13

3.2.2 推荐系统架构及推荐引擎 ……………………………………………………14

4 母婴用品推荐系统的设计与算法 …………………………………15

4.1 母婴用品推荐系统的前言 ……………………………………………………………15

4.2 母婴用品推荐系统的构成设计 ………………………………………………………15

4.1.1母婴用品推荐系统的功能模块 …………………………………………………15

4.1.2母婴用品推荐系统的工作流程 …………………………………………………16

4.3 母婴用品推荐系统的算法 ……………………………………………………………17

4.2.1 母婴用品推荐系统推荐引擎的架构 …………………………………………17

4.2.2 母婴用品推荐系统推荐算法的运行过程 ……………………………………18

5 母婴用品推荐系统的模拟实现 ……………………………………20

5.1 母婴用品推荐系统实现的编程工具介绍 ……………………………………………20

5.2 母婴用品推荐系统实现流程 …………… …………………………………………21

5.2.1 母婴用品推荐系统的前端工作 ………………………………………………21

5.2.1 母婴用品推荐系统的后台工作 ………………………………………………21

6 课题总结与改进 ……………………………………………………22

参考文献 ………………………………………………………………23

致谢 ……………………………………………………………………24

定制化母婴用品推荐系统

毕兴华

，China

Abstract：18 the fifth plenary session decided to open up a second child policy. This policy has been implemented formally since January 1, 2016 .It means that one couple can give birth to two children.This policy will bring China 8 million newborns in the next five years, bringing about a wave of baby boomers.This will be a huge market and business opportunities, because it is behind the surge in demand for maternal and infant supplies.

China's current internet users has reached 750 million, which mainly deposed of young adults.They have stronger curiosity to new things and the ability to adapt them and they also have stronger consuming desire and power.This reality combined with the upcoming baby boomers having a huge commercial value.Because that means the explosive growth of the demand for maternal and infant supplies.This study combined with e-commerce, maternal and child supplies recommendation system has huge commercial prospects.

This paper based on the basic reality that the maternal and infant supplies’ number is numerous and complicated.For example ,there are thousands of brands of milk powder in front of you .Now the biggest problem is how to select the user's most favorite or most interested ones.Besides, the products’ price, quality, composition must satisfy users’ requirement.The last problem is how to realize this function ,how to remove redundant amounts of information and how to produce the most appropriate maternal and infant supplies recommendation list.

The core of this paper is hybrid recommendation algorithm, which adopts assembly line type classification of hybrid algorithm.The first step is to use Content-based Recommendation Algorithm by combining the character of the products and the users’ interest to produce a set of vector based on user and product model.And then use the cosine similarity formula to calculate each product’s recommendation level.Finally,it produces the Top-N recommendation list.But the first algorithm wouldn't recommend these products which is determined by recommendation level to our users in the recommended moment .The second step is to use collaborative filtering algorithm, this algorithm determines the closest user by accounting the similarity of the first algorithm’s output .Finally,it recommends and the similar users’ maternal and infant supplies.This theoretical framework based on users’ record ,context parameters and product features ,which is a kind of optimized algorithm.In this model,the first recommendation system’s output will becomes the second recommendation system’s input.The two recommendation systems structurally constitute a bigger recommendation system.So the bigger system can produces the most reasonable maternal and infant supplies,such as milk powder and so on.It can also make sure that the recommend products’ price and quality could satisfy our recommendation system’s users.

Key words：Maternal and child supplies；Hybrid recommendation algorithm；Hierarchical hybrid；

剩余内容已隐藏，请支付后下载全文，论文总字数：23663字

相关图片展示：