The Contextual Group Recommendation
Jing Wang
1
, Hui Li
2
School of Computer Science and Technology
Xidian University
Xi’an, China
1
wangjing@mail.xidian.edu.cn
2
lihui@mail.xidian.edu.cn
Hui Zhao
School of Electronic & Information Engineering
Xi’an Jiaotong University
Xi’an, China
huizhao@stu.xjtu.edu.cn
Abstract—The context-aware calculation method is applied to
the model of recommendation system in this paper. Based on
context of scenario analysis and inference, we can implement
recommendation for user group. This paper makes use of the
context of scenarios aware technology, which fuses multi-
source data from smart phones, sensor networks and social
networks for the scenario modeling and inference, and then
formats the context clues to build the perceptual model of the
context scenarios to predict the behavior patterns of the user
group for the group recommendation, which meet the needs of
the user group. We have analyzed the recommendation process
and performed a subjective test to show the usefulness of the
proposed system.
Keywords-context-aware; sensor networks ; social networks;
group recommendation
I. INTRODUCTION
With the development of aware communication
technology and mobile communications technology, more
and more diverse information is needed. How to provide
information services to anyone at anytime and anywhere
about anything conveniently and effectively using mobile
devices becomes an urgent problem needed to address now.
In this paper, we combine context-aware information and
personalized behavior to form a context-aware
recommendation system. The former is used to get and
represent information data in any way at anytime and
anywhere and the latter can help us to discover what the
users need in huge behavior data and understanding what
they are doing and what they want to do based on their
context.
Actually, when you are making a decision, you will be
influenced by other people surrounding you obviously. This
is usually referred to as emotional contagion, and this
infection is usually proportional to the level of trust between
the people. The more you trust someone, the greater impact
on your. So, we take social factors into account, such as
individual’s personality and the density relationship between
friends, which can improve the prediction accuracy of the
group recommendation system. The group recommendation
generates a polymerization beyond the individual’s
preferences model to meet the preferences of all users in the
group [1, 2].
At the same time, in different states or scenarios, user’s
preferences of the same things will be different. For example,
when we are in a relatively quiet state may prefer soothing
music, noisy environment like a relatively passionate music.
In this paper, we use mobile devices, a variety of sensors
and social networks to get the user’s variety of information
data. Then, combine the affect of the social relationship and
scenario of the users, we proposed a context scenario
semantics-aware group recommended model, which achieves
group recommended by context of scenarios analysis and
inference.
II. R
ELATE WORKS
A. Context-aware Recommendation System
There are already some studies about context-aware
recommendation system [3]. Reference [4] studied the
impact of context on customer purchasing behavior, as well
as the trust relationship between users to create a context
model. Based on the entire context model, they achieve
recommendations related to online shopping.
Reference [5] proposed the decomposition machine
(FMS) to model contextual information and to provide
context-aware rating prediction. The FM model equations
can calculate the text variable in linear time and the size of
the decomposition, applicable to many different scenarios
and higher forecast, to solve the complexity of the linear
model. This method can quickly get the context of the
recommendations.
Reference [6] presented a personalization algorithm for
recommendation which relies on hierarchical tag clusters.
The recommendation framework is independent of the
clustering method, but we use a context-dependent variant
of hierarchical agglomerative clustering which takes into
account the user’s current navigation context in cluster
selection.
B. Group Recommendation
With the emergence of group behavior, there are some
studies for entire group recommendation. Reference [7]
proposed an algorithm to recommend appropriate and novel
content to groups of people. Based on the Power Balance
Map and the Behavioral Tendency of each group, the
algorithm recommends new content in or near high-density
areas on the group’s feature space which consists of
2013 5th International Conference on Intelligent Networking and Collaborative Systems
978-0-7695-4988-0/13 $26.00 © 2013 IEEE
DOI 10.1109/INCoS.2013.27
127
下载后可阅读完整内容，剩余4页未读，立即下载
weixin_38596117
- 粉丝: 12
- 资源: 913
我的内容管理
展开
最新资源
- JavaScript DOM事件处理实战示例
- 全新JDK 1.8.122版本安装包下载指南
- Python实现《点燃你温暖我》爱心代码指南
- 创新后轮驱动技术的电动三轮车介绍
- GPT系列：AI算法模型发展的终极方向？
- 3dsmax批量渲染技巧与VR5插件兼容性
- 3DsMAX破碎效果插件：打造逼真碎片动画
- 掌握最简GPT模型：Andrej Karpathy带你走进AI新时代
- 深入解析XGBOOST在回归预测中的应用
- 深度解析机器学习：原理、算法与应用
- 360智脑企业内测开启，探索人工智能新场景应用
- 3dsmax墙砖地砖插件应用与特性解析
- 微软GPT-4助力大模型指令微调与性能提升
- OpenSARUrban-1200：平衡类别数据集助力算法评估
- SQLAlchemy 1.4.39 版本特性分析与应用
- 高颜值简约个人简历模版分享
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
