SNS Information Extraction for Social Search

SNS Information Extraction for Social Search
Soo- Jeong Choi
Ok Ran Jeong, Ph.D.
Dept. of Software Design & Management
Gachon University
Seongnam, Republic of Korea
[email protected]
Dept. of Software Design & Management
Gachon University
Seongnam, Republic of Korea
[email protected]
Abstract— Many websites providing the SNS are appearing
due to the development of the web and the development of
various networks and internet technology to follow. In this paper,
we propose a method to extract the social information of users
from the social network websites followed by the promotion of
SNS. The social information of users would be able to provide the
information which users need if used for search or
recommendation system. In order to secure the social network
information of users from the SNS websites, the Open API
provided by SNS websites has been used.
Keywords—SNS(social network service, social information,
social search)
I.
INTRODUCTION
The web based digital industry is playing an important role
of economy while the social network based digital media and
the contents to follow are getting explosively increased under
the next general technology environment.[1] The information
desired by users requires the social network related information
connected to themselves and anticipates an individualized
custom-made search result which is differentiated from other
search websites rather than simply producing the search result.
Such social search requires the SNS information of users. This
paper attempts to propose a method to extract the SNS
information of users from the social network websites in order
to perform a social search. The 3 types of SNS websites mostly
widely used among people will be used to obtain user
information. These 3 types of SNS websites are Me2day,
Twitter and Facebook. The Open API provided by each
website will be used in order to extract the information of users
from these SNS websites.
The remainder of this paper is organized as follows. In
Section 2, we describe the extraction method of SNS
information on the search system. In Section 3, we describe the
preliminary experiment and results. In Section 4, we conclude
the paper.
II.
EXTRACTION METHOD OF SNS INFORMATION
extraction method has used the Open API provided by the
SNS websites while the actually applied websites are
Facebook, Twitter and Me2day. The detailed extraction
method is as follows.
A. Facebook API
Facebook provides an Open API according to the
JavaScript, PHP, mobile platform and language. Not only that,
but Facebook has also included the Graph API as a platform
component in order to connect up to a third party external
webpage. The open data in Facebook can be freely read
through this Graph API and the data can be written on the
Facebook platform using this Graph API. Using this, the user
information and friends list are extracted. The imported method
is as follows.[3]
https://graph.facebook.com/”User ID”/friends?access_token=”token key”
The return value gained by entering the url above is shown
as a form of JSON object. JSON has a format such as { "data":
[ { "name": "Jiyun Yoon", "id": "690349501" } ] } and the
friends list can be extracted using JSON Parsing. JSON is
simply a method of displaying data. It has a fast parsing as the
one created for a little simpler expression than the XML. It also
has advantage of being useful under mobile and web
environment.
Facebook has the Access Token which is a security device
to prevent from easily taking away the user information unlike
Me2day or Twitter. A unique Access Token must be issued for
each user to import the user information. Therefore, an App
must be registered to the Facebook using an authenticated
Facebook ID that guarantees identity such as phone number or
credit card to issue the Access Token that we want. If an App is
registered, a unique ID and unique Secret Code on the App can
be issued to get access from the user who has allowed this App.
If the user allows this App, an Access Token gets issued to this
user and the authorized information can be imported using this
Token.
In order to apply the social information on the search
system for users, this study has extracted the social
information related to the communities forming relationship
with the user according to the social network (writings,
comments, friends, user information and writing link). The
978-1-4799-0604-8/13/$31.00 ©2013 IEEE
But the Facebook has set a limit on the Access Token
expiration period and there is a problem of having to get the
Access Token issued again after going through the above
procedure if the expired.
B. Twitter API
Twitter has a provided library unlike other websites. This
library can be used in Java under the name of Twitter4j. This
can be used to import user information. However, there may be
a difficulty in importing the user information from having to
get authentication on each individual as this method requires
Twitter authentication. Therefore, the method found in place of
this one is the method such as Me2day explained above. This
method is as follows.[2]
http://api.twitter.com/1/statuses/friends.xml?screen_name=”User ID”
The method like above is the one using the GET method.
Information such as friends or timeline can be imported using
this method. The information that has been imported is shown
as XML format and extracts information using the XML SAX
Parser that had been used during the Me2day Parsing. While
parsing in Twitter, the user information must be imported with
screen_name instead of user ID.
C. Me2day API
The me2API request of Me2day can be used by accessing
data and functions provided by me2day. The request of
me2API uses GET or POST method using the HTTP request
method. As the requesting method of Among these, only user
information and friend related information are extracted. The
importing method is as follows.[1]
http://me2day.net/api/get_friends/”User ID”
If the me2API is requested by inserting the user ID you
wish to import to the “user ID” using the method above, you
can see user information and the information on the friends of
user. While the response of me2API could be seen as XML
format and JSON format, it is received as XML formation
among these two types of instances. If called as JSON format,
the result can be returned by assigning a callback parameter.
(Ex: json?callback=my_func”User ID”.json?callback_myfunc)
The required user information and friends list are extracted
on the returned result using XML SAX Parser. In this process,
SAX Parser is used as the parser. The reason is because SAX
Parser is more effect than DOM Parser in a situation of having
the read the entire XML. While parsing, the SAX Parser parses
in a series as an event is created each time the nodes come in
unlike the DOM Parser which creates as tree structure. Since
SAX Parser parses in a series, the reading speed is fast and
requires little memory. Due to such advantage, SAXParser is
selected to extract the Me2day information.
III.
PRELIMINARY EXPERIMENT AND RESULT
The preliminary experiment of this paper was performed
based on the Facebook. The reason is because there is a
security device to prevent importing user information easily in
case of Facebook although the basic information of user can be
imported easily in case of Me2day or Twitter. This security
device had a problem of having to get authentication again in
case of being used for a long time as it is made of Access
Token and the period is set up. Therefore, the method of being
able to use the Access Token continuously has been
experimented.
It has been made to receive the Access Token automatically
if the user is using our website for the first time. The default
value of this Access Token is set as 1 hour. In case the Access
Token is expired, it has been made to return as 60 day token
after going through ID and Secret Code authentication of the
App registered in advance using the 1 hour token already
issued. Using this, experiment has been performed with the
method extending the Token automatically limited to the
Tokens that almost expired by managing the tokens issued at
the Database. Through this method, the Token that can be used
continuously can be obtained if the user goes through the
approval on the authority just once at the beginning.
IV.
CONCLUSION
The SNS information for social search has extracted the
basic information of user and the friends list of user. This is
used while informing the search result and can be used for the
recommendation algorithm to inform the appropriate
recommendation result for each user. Also, if the weight is
calculated using SNS information and this is applied to the
proper recommendation algorithm, the result including social
factors can be shown. Such differentiated search result is able
to arouse interest of users.
ACKNOWLEDGMENT
This work was supported by Gachon University Research
Fund in 2013, and by Basic Science Research Program through
the National Research Foundation of Korea (NRF) funded by
the Ministry of Education, Science and Technology (20120004177)
REFERENCES
[1]
[2]
[3]
[4]
G. Lawton, “Knowledge Management: Ready for Prime Time?, IEEE
Computer, Vol. 34, No.2, pp216-244, 2001
Facebook
Developers
:
http://developers.facebook.com/docs/reference/api/
Twitter Developers : https://dev.twitter.com/
Naver
Developers
Center:
http://dev.naver.com/openapi/apis/me2day/me2api_intro