ABSTARCT interested in and relying on social network for

ABSTARCT

            Social network has gained remarkable attention
in the last decade. Accessing social network sites such as Twitter, Facebook, LinkedIn
and Google through the Internet has become more affordable. People are becoming
more interested in and relying on social network for information, news and
opinion of other users on diverse subject matters. Data mining provides a wide
range of techniques for detecting useful knowledge from massive datasets like
trends, patterns and rules. Data mining techniques are used for information
retrieval, statistical modelling and machine learning. These techniques employ data pre-processing,
data analysis and data interpretation processes in the
course of data analysis. This survey discusses different data mining techniques
used in mining diverse aspects of the social network over decades going from
the historical techniques to the up-to-date models.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

 

KEYWORDS

          Stress,
Information Retrieval, Statistical Modeling, Machine Learning, Spam Filtering, Online
Social Network (OSN), Phishing

 

INTRODUCTION

          The widespread use of
social media may provide opportunities to help reduce undiagnosed mental
illness. A growing number of studies examine mental health within social media
contexts, linking social media use and behavioral patterns with stress,
anxiety, depression and other mental illnesses. The greatest number of studies
of this kind focuses on stress. Social network is a graph consisting of nodes
and links used to represent social relations on social network
sites. The nodes include entities and the relationships between them
forms the links Social networks are important sources of online
interactions and contents sharing  subjectivity, assessments, approaches,
evaluation, influences, observations 
feelings, opinions and sentiments expression borne out in text, reviews,
blogs, discussions, news, remarks, reactions, or some other documents .

               
                                                                                                                                         

Edges

 

 

Fig 1: Interaction of users in
social network

 

          Recent research work
reveals that the stressed users are less interactive on social activities than
the non-stressed users.

 

 

Fig 2: Probability of interaction
of users in social media

 

ANALYSIS
ON SOCIAL NETWORK

            According to Technorati, about
75,000 new blogs and 1.2 million new posts giving opinion on products and services
are generated every day. Also massive data generated every minute on common
social network sites are laden with opinion of users as regards diverse subject
ranging from personal to global issues. User’s opinions on social network sites
can be referred to as discovery and recognition of positive or negative
expression on diverse subject matters of interest. These opinions are often
convincing and their indicators can be used as motivation when making choices
and decisions on patronage of certain products and services or even endorsement
of political candidate during elections. Even though online opinions can be
discovered using traditional methods, this form is conversely inadequate
considering the large volume of information generated on social network sites.
This fact underscores the relevance of data mining techniques in mining opinion
expressed on social network site.

Data
mining tools already used for opinion and sentiment analysis include collections
of simple counting methods to machine learning. Categorizing opinion-based text
using binary distinction of positive against negative is found to be
insufficient when ranking items in terms of recommendation or comparison of
several reviewers’ opinions. Data mining techniques used for opinion mining on
social network.

 

ANALYSING
STRESS USING SENSOR AND NEURAL NETWORK

            A range of
non-textual methods to passively detect stress have been developed. Stress cues
can be detected in the sound waves of people talking and this could become a
standard tool within smart phones. A more intrusive approach with sensors to
detect bodily responses like sweating or heart rate changes in order to detect
specific instances of stress during the day or during a commute. Within social
media, a neural network approach has been used to detect short term or longer
term stress in social media for individuals, based upon a range of factors,
including the emotions expressed in their tweets. Stress may also be passively
inferred from mobile phone call activity patterns and from oxygen levels in
exhaled breath, measured at a distance. Long term psychological states, such as
post-traumatic stress disorder have also been estimated from social media. It
is possible to detect depression from the profiles of users before its medical
diagnosis, through an increased uses of negative sentiment and tight online relationships.

A
practical application of this might be in the early detection of mothers that
are at risk of post-natal depression. There have been some attempts to detect
stress through language, although these have typically focused on long term
health-related stress rather than stressors or short term stress. The program
Linguistic Enquiry and Word Count (LIWC) counts the occurrences of a set of categorized
terms (e.g., positive emotion words, pronouns) within texts written by
individuals in order to identify patterns that will help to detect
psychological differences, such as for diagnosing psychiatric conditions. This
approach found that the use of more negative words by people writing about a
traumatic experience predicts a future lack of health improvements.
Post-traumatic stress disorder has also been identified through categories of
LIWC terms that tend to be used more often by self-declared sufferers in
Twitter than by other random users. Hence, it is clear that it is reasonable to
analyze the text written by people in order to detect long term stress,
although there is no equivalent evidence about short term stress.

 

 

 

 

Fig 3:  Respiration based stress assessment

 

STRESS
DETECTION USING CLUSTERING TECHNIQUES

Opinion
of influencers on social network is based largely on their personal views and
cannot be hold as absolute fact. However, their opinions are capable of
affecting the decisions of other users on diverse subject matters. Opinions of
influential users on Social network often count, resulting in opinion formation
evolvement. Clustering technique of data mining can be utilized to model
opinion formation by way of assessing the affected nodes and unaffected nodes.
Users that depict the same opinion are linked under the same nodes and those
with opposing opinion are linked in other nodes. This concept is referred to as
homophily in social network. Homophily can also be demonstrated using other
criteria such as race and gender.

 

 

Fig 4: Clustering

 

 

STRESS
DETECTION BASED ON CHRONOCITY

Stress
detection is done by combining Open Source Intelligence (OSINT) and
user-generated content classification techniques with a user-driven stress test
as applied to a community of Online Social Network (OSN) users. The chronocity
of the users is detected. Chronocity is the deviation the user’s interaction in
the social media. Unsupervised flat data classification of the user-generated
content is performed to formulate two working clusters which classify usage
patterns that depict medium-to-low and medium-to-high stress levels
respectively.

 

Table
1: Different Approaches to Detect Stress

 

 
APPROACH

 
TOOLS

 
DESCRIPTION

Graph
Theoretic

Centrality
measure

Inspects
representation of
power
and influence that forms clusters and cohesiveness

a-centrality

Measures
the number of alleviated paths that exist among nodes

Clustering

Structural
equivalence
Measures

Detects
friendship structure on social network based on
shared
behavior

Opinion
Analysis

Aspect-
Based/Feature-
Based

To
identify positive or negative opinion sentences in product reviews

Opinion
Formation

Homophily
Clustering

Used
to link same opinion under the same nodes

Opinion
Definition and Opinion
Summarization

Support
Vector
Machine

Used
to learn the polarity of neutral examples in documents

Unsupervised
Classification

Part
of speech tagging

Used
to rate a review as ‘thumbs up’ or ‘thumbs down’

 

ONLINE
SPAM FILTERING IN SOCIAL NETWORKS

Online spam filtering system is used to inspect messages in
online social networks. Text shingling and incremental clustering are used to
reconstruct spam messages into campaigns in real-time for classification rather
than to examine them individually. The system adopts novel features that
effectively characterize spam campaigns. The challenge of reconstructing
campaigns in real-time is done by adopting incremental clustering and
parallelization.

            The below diagram depicts that
online spam filtering has been deployed in Online Social Network (OSN). Every
message is inspected before rendering the message to the intended recipients
and immediate decision on whether the message under inspection should be
dropped or not is made. The below diagram depicts that the spam message
originates from a compromised account.

Fig 5 : Online Spam Filtering

SPAMMERS
ON TWITTER

            The detection of
spammers on twitter is done by collecting a large dataset of Twitter that
included more than 54 million users, 1.9 billion links, and almost 1.8 billion
tweets. The tweets were manually classified into spammers and non-spammers. A
number of characteristics related to tweet content and user social behavior
were identified, which could potentially be used to detect spammers. The
characteristics as attributes of machine learning process for classifying users
as either spammers or non spammers were used. This strategy succeeded at
detecting much of the spammers while only a small percentage of non-spammers
are misclassified. Approximately 70% of spammers and 96% of non-spammers were
correctly classified.

 

 

Fig 6: Social Spam on Twitter

 

 

HONEYPOTS
FOR DETETCION OF SPAMMERS IN SOCIAL MEDIA

Detection
of spammers is done using the concept of honey pots. It
is used for harvesting deceptive spam profiles from social networking
communities. It plays a major part in statistical analysis of the properties of
these spam profiles for creating spam classifiers to actively filter out
existing and new spammers.  It is found
that the deployed social honey pots identified social spammers with low false
positive rates and that the harvested spam data contains signals that are
strongly correlated with observable profile features (e.g., content, friend
information, posting patterns, etc.). Based on these profile features, machine
learning based classifiers  was
developed, which was used for identifying previously unknown spammers with high
precision and a low rate of false positives.

 

SOCIAL PHISHING

            Phishing is a network type attack where the
attacker creates the fake of an existing webpage to fool an online user into elicits
personal Information. Phishing is the combination of social engineering and
technical methods to convince the user to reveal their personal data. It is
typically carried out by Email spoofing or instant messaging. A novel anti-phishing approach based on Dynamic watermarking
technique is used to detect spammers.

 

REFERENCES

 

1      Huijie Lin, Jia
Jia, Jiezhong Qiu, Yongfeng Zhang, Guangyao Shen, Lexing Xie, Jie Tang, Ling
Feng, and       

          Tat-Seng Chua , 2017
,Detecting Stress Based on Social Interactions in Social Networks.

     2      Manuel
Egele, Gianluca Stringhini, Christopher Kruegel and Giovanni Vigna , 2017, Towards
Detecting

                Compromised
Accounts on Social Networks.

     3   
  Miltiadis Kandias,Dimitris Gritzalis, Kostas
Nikoloulis, Vasilis Stavrou, 2017, Stress level detection via                           OSN usage pattern and chronocity analysis: An
OSINT threat intelligence module.

      4      Nandita Sharma, Tom Gedeon , 2017, Objective
measures, sensors and techniques for stress recognition a

                 Classification.

     5     Mike Thelwall, 2017, TensiStrength: Stress
and relaxation magnitude detection for social media text.

     6   
Sharath Chandra Guntuku1, David B Yaden1, Margaret L Kern – Detecting
depression and mental illness  

              on social
media: an integrative review.

     7    
F. Benevenuto, G. Magno, T. Rodrigues, and
V. Almeida , 2010, Detecting Spammers on Twitter in  

              Conference on Email and Anti-Spam
(CEAS).

    
8    H. Gao, Y. Chen, K. Lee, D.
Palsetia, and A. Choudhary, 2012 , Towards Online Spam Filtering in Social  

              Networks in Symposium on Network
and Distributed System Security (NDSS).

    
9    K. Lee, J. Caverlee,
and S. Webb,2010,Uncovering social spammers: social honey pots + machine  

              learning, in International ACM SIGIR
Conference on Research and Development in Information

              Retrieval.

   
10   C. Grier, K. Thomas, V.
Paxson, and M. Zhang,2010,@spam: the underground on 140 characters or less, in    

             ACM Conference on Computer and Communications
Security (CCS).

   
11  T. Jagatic, N. Johnson, M.
Jakobsson, and T. Jagatif,,2007,Social Phishing, Comm. ACM, vol. 50, no. 10,     

             pp. 94–100.

x

Hi!
I'm Mack!

Would you like to get a custom essay? How about receiving a customized one?

Check it out