《个人知识管理平台下共享资源的个性化推荐.doc》由会员分享，可在线阅读，更多相关《个人知识管理平台下共享资源的个性化推荐.doc（71页珍藏版）》请在金锄头文库上搜索。
1、分类号 TP311.1 密 级 U D C 编 号 10486 硕 士 学 位 论 文个人知识管理平台下共享资源的个性化推荐： 学 号： 指导教师姓名、职称： 教授：计算机软件与理论：软件工程二一二年五月郑 重 声 明本人的学位论文是在导师指导下独立撰写并完成的，学位论文没有剽窃、抄袭、造假等违反学术道德、学术规范和侵权行为，否则，本人愿意承担由此而产生的法律责任和法律后果，特此郑重声明。 学位论文作者（签名）： 年 月 日摘要随着互联网的发展和网络资源的急速增长，为了满足人们对于个人知识、信息进行管理的需求，个人知识管理平台应运而生。同时为满足人们随时随地在任意平台下均可进行个人知识管理的需
2、求，出现了云笔记。云笔记下的海量信息，是用户经过加工处理过的知识，若将其共享出来，实现知识价值最大化的同时达到用户之间信息共享的目的。与此同时，信息定位困难的问题随之出现，因此通过个性化推荐解决信息定位困难的问题具有重要的研究意义和现实价值。本文依托东方网景公司和实验室合作的云存储个人知识管理项目，通过分析云端个人知识管理平台的系统架构、逻辑层次及工作流程，以及结合个人知识管理平台的特点，提出了将云笔记下个人的知识资源在云端实现共享的思想。同时针对资源共享带来的海量信息造成的信息定位困难的问题，提出了对其进行个性化推荐的需求，即针对平台客户端用户进行个性化资源推荐和相似好友推荐。通过分析个人知
3、识管理客户端存储用户知识、笔记的特点，设计实现了共享资源的个性化推荐服务。本文个性化推荐的工作主要包括：在个性化推荐需求分析的基础上，设计个性化推荐服务的框架，分层为：数据处理层、推荐层和用户交互层，推荐流程设计并详细设计三层模式中的各功能模块。推荐模块基于协同过滤的推荐策略下产生候选资源，将预测评分排序，Top-N来进行资源的推荐。用户聚类通过计算用户相似性，实现对用户好友的推荐及社区的形成，从而实现资源更好的共享和交流平台。本论文实现了个性化推荐的核心推荐模块，并基于实验数据集设计了实验来验证本文设计思路的可行性，并且实验结果表明在个人知识管理平台下共享资源的个性化推荐具有可行性和实际意义
4、。关键词：个人知识管理系统 个性化推荐 内容推荐策略 协同过滤推荐IIABSTRACTWith the rapid growth of the Internet and network resources, personal knowledge management platform has emerged to satisfy the needs of personal knowledge and information management. Meanwhile the emergence of cloud notes can satisfy the demand for using pe
5、rsonal knowledge management in any platform .Large amounts of information of cloud notes come from the modified knowledge by the users. Resources sharing can maximize the value of knowledge and achieve the purpose of resources sharing among users. However，the information location is too difficult to
6、 realize resources sharing. Therefore，information personalized recommendation is proposed to solve the difficult problems which indicated to be of great value.This paper relies on the cloud storage of personal knowledge management projects by the cooperation of East Netscape and laboratory. On the b
7、asis of the analysis of cloud personal knowledge management platform, system architecture, the logic level and workflow, as well as the characteristics of the combination of personal knowledge management platform, the design of this paper provides an ideology sharing of the personal knowledge resour
8、ces in Cloud notes. The difficulties of position of information caused by the mass of information give rise to resource sharing which on the basis of personalized recommendation platform and client user personalized resource recommendation as well as similar friends.The design is in the service of p
9、ersonal recommendation based on sharing resource by analyzing personal knowledge management of client storage users as well as the characteristics of notes. The main work content consist of the frame design based on custom designed demand which is subdivided as data processing, recommendation layer
10、and client layer. Whats more, process recommendation design and three-layer mode as well as functional module are also included. On the basis of content and recommendation strategy, the recommend module gives rise to candidate resources, prediction score, Top-N, thus the resource recommendation take
11、s shape. The users clustering realizes user recommendation and communities by computing the similarity of the users. Consequently, there comes into being a platform for resource sharing and communication.After all, the paper has realized the key module of resources personalized recommendation, whats
12、 more, the feasibility of the results have been tested by experimental data sets which accordingly reveal that resources personalized recommendation indeed does good to the personal knowledge resources management and possess the valuable significance. Keywords: personal knowledge management system,
13、personalized recommender, Content-Based recommended strategy , Collaborative Filtering目 录摘要IABSTRACTI1引言11.1研究背景及现实意义11.2国内外研究现状21.3论文研究内容41.4论文研究目标51.5论文研究结构52个人知识管理和个性化推荐概述72.1个人知识管理的概念72.1.1知识的定义72.1.2个人知识管理的定义72.1.3个人知识管理的原则82.1.4个人知识管理的目的92.2个人知识管理工具102.2.1常用个人知识管理工具102.2.2个人知识管理工具缺点122.3个性化推荐1
14、32.3.1个性化推荐132.3.2个性化推荐相关术语132.3.3个性化推荐流程142.4个性化推荐技术152.4.1基于内容的推荐(Content-Based)172.4.2协同推荐(Collaborative)192.4.3混合推荐(Hybrid)242.5本章小结243个人知识管理平台介绍及个性化推荐需求分析253.1平台总体部署253.2系统逻辑层次263.3系统工作流程273.4个性化推荐需求设计283.4.1共享资源的原则293.4.2推荐的主体293.4.3推荐的内容303.5本章小结304个性化推荐服务设计314.1个性化推荐服务的设计目标314.2个性化推荐服务整体框架31
15、4.2.1个性化推荐服务框架314.2.2个性化推荐流程324.3数据处理层的设计344.3.1兴趣建模模块344.3.2兴趣变化检测更新模块364.3.3用户聚类模块374.4推荐处理层设计384.4.1好友推荐模块384.4.2资源推荐模块394.5交互层设计414.5.1笔记操作模块424.5.2界面交互模块454.6个性化推荐数据库表设计464.7本章小结495实验系统505.1实验环境505.2实验过程515.2.1实验数据集515.2.2数据预处理525.2.3相似用户计算545.2.4项目推荐555.3实验结果及分析555.3.1常用推荐评估指标555.3.2结果及分析565.4本章小结586总结与展望596.1总结596.2展望59参考文献
