2. Information Retrieval Topics In Twitter Using Weighted Prediction Network
http://www.iaeme.com/IJCIET/index.asp 782 email@example.com
are also facilitated by weightage. This means user will get the priorities tags, topics and related users;
according to the weightage.
The suggested prediction network not only analyze an information on social networking sites but
also maps valuable tags and person to associate a user with invaluable information in the form of
2. LITREATURE SURVEY
Social networking is a platform where the entire persons like to spend the time to share information;
achievements in life, opinion on some topic can be expressed. In short it is the comfortable form of
There are number of social networking platform for the communication, two survey carried out at
Australia shows that 52% of usage of an internet has been increased due to social networking sites.
Total amount of Overall access of an internet due to social networking site is 79%. Also 49% of
internet is daily used in comparison of overall internet usage per day.
There are 45% people who use social networking site on an average 297 persons are connected
with every person. 70% mobile holders use an internet in the form of social networking sites.
Three social networking sites are successful to engage people on an internet, namely: Twitter,
Facebook, and LinkedIn. Statistics collected in year 2013 says that on an average 118 users follow a
count and 52% of total number of users tweets at least once in a week.
The relation, association are mentioned by using Graph, so one basic concepts needed to do the
social networking analysis using graphs are as below:
1. Vertices: It could be represented with user.
2. Edge: Edge is the connection among nodes. Here Edge could be used to represent weight.
There is some important measurement applied to know the association, relation and following
metrics are used.
1. Density of network: If there are more number of ties in a graph, then graph is known as Dense
graph. The graph which contains less number of ties are known as spars graphs.
2. Betweenanes: The route with the help of which two nodes are connected directly. In case of social
networking analysis, these nodes could be person or trend.
3. Closeness: This is the path through which one node is connected with another node.
There are different types of networking with various structures. According to the nature of the
graph the analysis of social networking analysis need to be made.
The example of a graph is mentioned below:
Figure 1 Simulated graph of association of data of social networking site – Twitter
3. Boshra F. Zopon AL_Bayaty
http://www.iaeme.com/IJCIET/index.asp 783 firstname.lastname@example.org
This is a graph G (V, E), where V indicates vertices A,B,C,D and E indicates edges AB, BC, CD,
DE, EF, FA, BF, BE, CF, CE. Here the node may represent object and edges mention relationship
amongst it. Simulation of social networking websites could be made with the help of a graph which
may help to analysis the relationship among the node.
There are number of approaches to suggest or search the relevant record based on the information
like graph. Members are suggested with the adequate relationship by considering the extent of
relationship as well.
There are different types of graphs present with and without weight. There are social networking
sites like twitter, the where the dummy records could be easily created and with the help of it analysis
could be done problem in case of existing structure is there is no prediction made to calculate interest.
Figure 2 A-graph without weight B: Graph with weight
3. PROPOSED ARCHITECTURE
Figure 3 Weighted prediction network for information retrieval from Twitter
Proposed architecture says that the social networking site helps to collect data from social
networking site. The collected data is stored in a database.
The stored data helps to analysis the data by the means of weighted prediction network, the trends
are analyzed and the best suitable data is revealed to the user. The important or significant change
4. Information Retrieval Topics In Twitter Using Weighted Prediction Network
http://www.iaeme.com/IJCIET/index.asp 784 email@example.com
could be observed in the form reduction in time span required to search similar or relevant record from
a same data source like twitter. This type of valuable information generated with the help of the data
generated by the model mentioned above. This optimized structure of the prediction network helps to
suggest the topics relate with the trends displayed on account.
Figure 4 Weighted framework for trends and tweets
The graph mentioned above gives detailed functioning of weighted prediction network. This graph
is nothing but the instance or snapshot of actual network.
Consider example mentioned above indicates the trends mentioned in a twitter. The logic used to
display the trend is a recent topic, but user may be interested in number of topics which may not be
mentioned in the list of trends. Also what sort of logic or technique to be utilized to display the trends
and their sequence.
In the graph mentioned above A….G, represents the various trends available on twitter.
Considering that A is a trend which is available; Most relevant trend is to be collected on every edge
the weights are available unlike of approach used in Markov chain model, where the node which is
generally achieved by using shortest path is calculated. This type of approach is useful when reach
ability or communication in a time is main concern. The aim of this experiment is to identify a trend
associated with current thread. Here ABCJGF is the path considering the relationship among node
with the higher weight out of all available node is considered as next trend. in this way the associated
node is identified as a the strongest associated node with a node A. In the graph mentioned above the
A is associated with AE, AC, and AB. In this association, AB>AE>AC, there AB association is
considered, so B node will get displayed as most relevant trend.
The input for this experiment is the credentials of twitter account; that is user need to have twitter
account. Once user Logged in into the system the list of recent trends is displayed with the help of
procedure mentioned above thus user will be explored only to the recent trends. With the help of
experiment performed the suggestion or the prediction of the most suitable trend is made to avail the
facilities or data source provided by the twitter.
Thus, the novel approach helps to display most relevant trends to the respective user as per the data
fetched by the system. In this application, the efforts required to search the relevant data is minimized
and this could be also verified or tallied by the amount of time span spent by the respective user on the
respective trend which is displayed trend which is displayed by weightage based prediction network.
5. Boshra F. Zopon AL_Bayaty
http://www.iaeme.com/IJCIET/index.asp 785 firstname.lastname@example.org
4. THE WEIGHTED PREDICTION ALGORITHM
Box: The weighted prediction algorithm
U--- User (with user’s credentials)
T--- Recent trends
Tw--- Twitter data source
TT--- Total number of trends
Twj, Twi --- Weightage of association among trend.
TLi --- List with descending order of weightage of trends.
5. EXPERIMENTAL SETUP
To perform this work, spring framework is used, this framework designed to facilitate the effective
and efficient programming by separating business applications. The data is stored in MySQL with the
help of the hibernate which maps object with database Eclipse Kepler is used, this is the version of
eclipse a IDE (Integrated development Environment) used develop applications in J2EE spring
framework. To execute the web applications designed using Java, Tomeat-7.0 server is used as web
6. IMPLEMENTATION AND TESTING
To perform the experiment spring, MVC architecture and framework is used. Advantage of this
technology is, one can modify view data or logic without affecting or bothering the associated
technologies. For that the spring framework is used. This is the web development for enterprise
(internet) using Java technology. Model view controller (MVC) approach is followed to perform this
experiment. Also, twitter APIs are used ensure communication with account and respective app on
twitter, this application program provide an interface to connect programs with different technologies.
Twitter account is needed along with API to fetch the information available on a twitter. Job of
prediction network is to understand the requirement of user and make the availability of information in
the form of tweets, trends and associated person (that is source of the information). The
implementation and testing face shown in screenshots below:
Step 1: Fetch <U, <T>>
Where U, <T> Tw
Step 2: Consider logical Graph of size TT
Where TT Tw
Step 3: For T1…TT
Where i 1… TT
Li = Ti
Step 4: For L1 … LT
6. Information Retrieval Topics In Twitter Using Weighted Prediction Network
http://www.iaeme.com/IJCIET/index.asp 786 email@example.com
Figure 5 Some screenshot related to implementation and testing face
Recent topics are called as a trend, these trends are fetched with the person who initiate and tweet on
the trends, snapshot of the data is mentioned below:
7. Boshra F. Zopon AL_Bayaty
http://www.iaeme.com/IJCIET/index.asp 787 firstname.lastname@example.org
Screenshot of Table -1 User – Recent trends
Social communication delivering useful data is segregated in adroit manner; structure of collecting
this data can be seen in a table mentioned below.
Screenshot of Table –2 Trends – Tweet s – user
8. Information Retrieval Topics In Twitter Using Weighted Prediction Network
http://www.iaeme.com/IJCIET/index.asp 788 email@example.com
Based on the work experiment performed the weighted prediction network helps to fetch the highly-
associated tweets from data source-twitter. This helps to effectively prioritize different tweets
available. Also, user can respond by inserting new tweets, which will get reflected to his/her original
twitter account. This work can be useful to minimize the information in the form of tweets.
Human nature is to share thoughts, pictures, Achievements with friends, colleague, and relatives.
Social networking site is a platform to achieve the publicity thus the huge amount of data becomes
source of information.
I would like to thank Al_mustansiriyah university, (www.uomstansiriyah.edu.iq), Baghdad, Iraq, for
its support in the present work and to inspire me always.
 Sensis, Social Media Report May 2015, Rob Wong, sensis,, The Digital Industry Association of
 Digital Marketing Analytics with a Competitive Edge, Social Media Trends 2015.
 Visualizing Big Data: Social Network Analysis, Michael Lieberman, Digital research conference,
 The use of Social Network Analysis in Innovation Research: A literature review Fabrice Coulon,
PhD Candidate1 Division of Innovation - LTH Lund University, Sweden, January 17th 2005.
 Shamanth Kumar Fred Morstatter Huan Liu Twitter Data Analytics, Springer, August 19, 2013.
 Sentiment Analysis of Twitter Data Apoorv Agarwal Boyi Xie Ilia Vovsha Owen Rambow
Rebecca Passonneau 2011.
 Massive Social Network Analysis: Mining Twitter for Social Good David Ediger Karl Jiang Jason
Riedy David A. Bader Georgia Institute of Technology Atlanta, GA, USA Courtney Corley Rob
Farber Pacific Northwest National Lab. Richland, WA, USA William N. Reynolds Least Squares
Software, Inc. Albuquerque, NM, USA, 2010 39th International Conference on Parallel Processing,
 Sentiment Knowledge Discovery in Twitter Streaming Data Albert Bifet and Eibe Frank University
of Waikato, Hamilton, New Zealand.
 Visual Sentiment Analysis on Twitter Data Streams Ming Hao, Christian Rohrdantz, Halldór
Janetzko, Umeshwar Dayal Daniel A. Keim, Lars-Erik Haug, Mei-Chun Hsu Hewlett-Packard Labs
and University of Konstanz.
 Boshra F. Zopon AL_Bayaty and Shashank Joshi, Empirical Implementation Naive Bayes
Classifier For Wsd, Using Wordnet. International Journal of Civil Engineering and Technology,
5(8), 2014, pp. 25–31.
 M.Rajeswari, Influence of Social Networking Sites on Personal and Professional Lives of People.
International Journal of Advanced Research in Engineering and Technology, 5(3), 2014, pp. 64–72.
 Nana Kwame Gyamfi, Makafui Nyamadi, Prince Appiah, Dr. Ferdinand Katsriku and Dr. Jama-
Deen Abdulai, A Brief Survey of Mobile Forensics Analysis on Social Networking Application.
International Journal of Computer Engineering and Technology, 7(4), 2016, pp. 81–86.
 How we analyzed Twitter social media networks with NodeXL, PEW research center, 2014.