Evaluating User Influence Based on Web2.0 UGC
Yue Zhai,Beijing University of Posts and Telecommunications, yiyimuduoduo@gmail.com,
Lei Li, Beijing University of Posts and Telecommunications, leili@bupt.edu.cn,
Xin Lin, Beijing University of Posts and Telecommunications, lx1989831@qq.com, and
Jiayin Qi,Beijing University of Posts and Telecommunications, qijiayin@139.com
Abstract—With the development of Social Networking
Services, users can publish and receive information
expediently. Social marketing emerged at this
circumstance with an important need to study the
influence of marketing activities in network. Based on the
research of user influence, choosing the most influential
users can improve marketing effectiveness greatly. With
the appearance of Facebook, Twitter, and other social
networks, more and more scholars began to study user
influence in such complex and huge network. In this paper,
we make a quantitative analysis to evaluate user influence
and classify users into different categories. We
synthetically consider various user attributes, including
the static social relationships and dynamic social activities.
We add new components to improve PageRank algorithm
and propose a behavior-based ranking(BBR) algorithm.
Experiments show that the improved BBR algorithm has a
satisfied performance.
Keywords—Social network, Analysis, PageRank, User
influence,BBR
I.
INTRODUCTION
UGC is User Generated Content; users publish their
original contents and supply them to other users through
Internet. Compared to Web1.0, Web 2.0 is very different.
Users obtain information through browser in Web 1.0, but Web
2.0 pays more attention to the interactions among users. Users
can build an environment which can be used to share, create,
and communicate and so on. We can find a lot of user
behaviors through researching on UGC. In this paper, we
collect various characters of UGC and use these information to
rank the user influence.
With the development of Social Networking Services,
people can publish and receive information expediently, which
leads to the fact which is online life affecting real life more and
more. Social marketing emerged at this circumstance. There is
a very important requirement which needs to make research on
the influence of marketing activity in network. So, analyzing
the user influence in network is very important.
In this paper, we choose Tianya forum as the research
object. Tianya forum is the most popular Internet forum in
China founded in March 1999. It provides BBS, blog,
microblog and photo album services. Up to now, in each month,
the covered quality users of Tianya are more than 200 million;
the registered users are more than 75 million. Tianya forum has
a great deal of high quality UGC; hence it can provide good
material for the research.
II.
RELATED WORK
In the circumstance of UGC, user influence has been
enhanced, and it has become the main driving force of the
development of UGC. As we can see, in some network
environment such as weibo and BBS, social networking is
mainly based on users and has formed a huge platform of users
and their interactions. Although ordinary users have relatively
small influence limited to the scope of their social scale and
social influence, authoritative users have much more influence.
So, for public opinion analysis or data mining and so on,
authoritative users have more obvious value for research. User
authority analysis, user ranking or leadership research finding
can be used to find more influential user.
Up to now, many scholars have dedicated to find
authoritative users. Since Sergey Brin and Lawrence Page
published PageRank
[1]
algorithm in Google and had a great
success, Link analysis algorithm has been widely introduced to
the tasks of user ranking. These algorithm are represented by
PageRank, Hits
[2]
, and SALSA
[3]
, etc. They mainly use the fans
and followers relationships as the links of authority
transmission between users.
Later research found that it is unilateral that we simply use
the static social relationship between the users as the basis to
evaluate user influence. Considering other useful information
can effectively improve the effect, make the results more
comprehensive.
Jian Jiao et al.
[4]
investigated the expertise that users
displayed in online communities, especially in discussion
groups and propose an effective expert ranking algorithm,
which integrates both discussion thread contents and social
network extracted from massive social interactions. They
presented a vector space model to compute the content
relevance part and a PageRank style algorithm for the expert
network part. Yamaguchi Y et al.
[5]
proposed TURank (Twitter
User Rank), which is an algorithm for evaluating users’
authority scores in Twitter based on link analysis. In TURank,
users and tweets are represented in a user-tweet graph which
models information flow, and ObjectRank is applied to
evaluate users’ authority scores.
These methods mainly analyzed based on link relations and
consider the more information to improve the rationality of
user ranking.