Prepared by Ashenafi Workie (pgr/18068/11)
Adama Science and Technology University
Department of Computer Science and Engineering
Critical Review Report on a course Design of Algorithm and Analysis
Submitted to Dr. Tilahun Melak (PHD)
Presentation Outline
2
Paper published in 2012 by
3
Abstract
This paper studied mining database of different algorithms.
like F-tree and Aprori
In this paper the authors tried to investigate forecasting
techniques based on directed weighted social graph.
In the methodology author’s research proposed is to
generate a social graph user’s action and predict the future.
Sanam et al. created some mathematical rules applied on
the graph in terms of community memberships
The strength of a relationship between two users without
knowing the content of the discussion.
4
Introduction
The article entitled with “Graph based forecasting for
social network site (Sanam Kadge and Gresha Bhatia)”
had been published at IEEE in 2012 Oct. 19-20, Mumbai,
India.
The authors study how to connect the people of similar
interest and builds some model.
They stated the an algorithms require more time for
scanning the datasets multiple times
multiple database scans for performing data mining on
the databases
5
Introduction cont...d
The basic idea of this research lies on forecasting based
on the generated social graph.
This novel approach for mining consists of visualization
of datasets into dynamic, directed, weighted graphs .
The graph depicts the user’s status in real time using
links and the weights on the links.
Lastly the authors introduce the limitation and overview
topic that had been included under the article contents.
6
Statements of the problem
As a result data is increasing day by day at a rate
which is difficult to control and process.
Even though researches have been carried out in this
area on huge datasets using association rule mining.
 These algorithms require more time for scanning the
datasets .
multiple times for example the algorithms require
multiple database scan for performing data mining.
7
Research Objective
Studied the approach of dynamic, directed, weighted
graphs and implemented Facebook like social network.
Specifically
Generate a social graph of user’s actions.
 predict the future social activities using graph mining
The strength of a relationship between two users
without knowing the content of the discussion.
8
Methodology(Data finding technique)
Authors stated that data can be collected from two
sources i.e.
1) The data entered by the user at the time of profile
creation and
2) based on users activities as follows like Login of sender
and receiver, searching a user, events like posting a
comment or liking a post, chatting
The methodology proposed is to generate a social graph
of users’ actions and predict the future social activities.
9
Methodology(Approaches)
1) Capture the data of user’s activities to generate graph
2) Forecasting using rule based mining.
10
Table 1. Strength of Relationship
Figure 1. Rule based mining
Implementation( social network screenshot)
11
Figure 2 and 3. implemented social networking site screen shot
Implementation framework analysis
12
Figure 4. implementation analysis
Result and discussion
Forecasting based on the result
The method is more efficient compared to Apriori
algorithm as there is no need of scanning the dataset
 Finding the accuracy of the predictions
 The real world simulation was done to find the
accuracy of the prediction made by the system.
 The predictions were noted on each day and after
observation for 10 days it was found that maximum
number of predictions turned true.
13
Cont.…d
14
Figure 5. Accuracy of predictions day wise
Figure 6:Accuracy of predictions percentage wise
Critique(strength of the paper)
1. The Authors(s) have appropriate professional
qualification and experience the main author studied
information technology whereas and the coauthor was
assistant professor(at department of computer
engineering ).
2. The abstract include all basic concepts and summary.
3. The abstract outlined methodology, reports of finding
limitation of the study.
4. The introduction stated overview the graph based
forecasting and the problems in paragraph1
15
Critique(strength of the paper)
1. The methodology section clearly state two approaches
their was captured data of user’s activities to generate graph
and forecasting using rule based mining.
2.the result section used graphics materials enhance clarity of
result through pie and other.
3. Generally, the data collection mechanism is the users
activity in the social networks
4.The conclusion state some how the basic results and
judgments to strengthen relationship.
16
Critique(Weakness of the paper)
1. The title of the research was too much
2. It is difficult to judge whether the authors of the
research had appropriate academic qualification
3. No recommendation were mentioned in the article.
4. The impact factors of the authors was not a good
because only 3 paper cited this paper for the last 7
years.
5. This paper major biased to research project or project
but it is not pure scientific research.
17
Conclusion
Generally, this paper good recommender system for
social networking friendship and community.
The result and methodology based on such forecasting
they could found the most popular community among
others
which would got maximum membership in the future
and in which the users are more active.
18
Any Question
19
Thank you

Graph based forcasting for social network

  • 1.
    Prepared by AshenafiWorkie (pgr/18068/11) Adama Science and Technology University Department of Computer Science and Engineering Critical Review Report on a course Design of Algorithm and Analysis Submitted to Dr. Tilahun Melak (PHD)
  • 2.
  • 3.
  • 4.
    Abstract This paper studiedmining database of different algorithms. like F-tree and Aprori In this paper the authors tried to investigate forecasting techniques based on directed weighted social graph. In the methodology author’s research proposed is to generate a social graph user’s action and predict the future. Sanam et al. created some mathematical rules applied on the graph in terms of community memberships The strength of a relationship between two users without knowing the content of the discussion. 4
  • 5.
    Introduction The article entitledwith “Graph based forecasting for social network site (Sanam Kadge and Gresha Bhatia)” had been published at IEEE in 2012 Oct. 19-20, Mumbai, India. The authors study how to connect the people of similar interest and builds some model. They stated the an algorithms require more time for scanning the datasets multiple times multiple database scans for performing data mining on the databases 5
  • 6.
    Introduction cont...d The basicidea of this research lies on forecasting based on the generated social graph. This novel approach for mining consists of visualization of datasets into dynamic, directed, weighted graphs . The graph depicts the user’s status in real time using links and the weights on the links. Lastly the authors introduce the limitation and overview topic that had been included under the article contents. 6
  • 7.
    Statements of theproblem As a result data is increasing day by day at a rate which is difficult to control and process. Even though researches have been carried out in this area on huge datasets using association rule mining.  These algorithms require more time for scanning the datasets . multiple times for example the algorithms require multiple database scan for performing data mining. 7
  • 8.
    Research Objective Studied theapproach of dynamic, directed, weighted graphs and implemented Facebook like social network. Specifically Generate a social graph of user’s actions.  predict the future social activities using graph mining The strength of a relationship between two users without knowing the content of the discussion. 8
  • 9.
    Methodology(Data finding technique) Authorsstated that data can be collected from two sources i.e. 1) The data entered by the user at the time of profile creation and 2) based on users activities as follows like Login of sender and receiver, searching a user, events like posting a comment or liking a post, chatting The methodology proposed is to generate a social graph of users’ actions and predict the future social activities. 9
  • 10.
    Methodology(Approaches) 1) Capture thedata of user’s activities to generate graph 2) Forecasting using rule based mining. 10 Table 1. Strength of Relationship Figure 1. Rule based mining
  • 11.
    Implementation( social networkscreenshot) 11 Figure 2 and 3. implemented social networking site screen shot
  • 12.
  • 13.
    Result and discussion Forecastingbased on the result The method is more efficient compared to Apriori algorithm as there is no need of scanning the dataset  Finding the accuracy of the predictions  The real world simulation was done to find the accuracy of the prediction made by the system.  The predictions were noted on each day and after observation for 10 days it was found that maximum number of predictions turned true. 13
  • 14.
    Cont.…d 14 Figure 5. Accuracyof predictions day wise Figure 6:Accuracy of predictions percentage wise
  • 15.
    Critique(strength of thepaper) 1. The Authors(s) have appropriate professional qualification and experience the main author studied information technology whereas and the coauthor was assistant professor(at department of computer engineering ). 2. The abstract include all basic concepts and summary. 3. The abstract outlined methodology, reports of finding limitation of the study. 4. The introduction stated overview the graph based forecasting and the problems in paragraph1 15
  • 16.
    Critique(strength of thepaper) 1. The methodology section clearly state two approaches their was captured data of user’s activities to generate graph and forecasting using rule based mining. 2.the result section used graphics materials enhance clarity of result through pie and other. 3. Generally, the data collection mechanism is the users activity in the social networks 4.The conclusion state some how the basic results and judgments to strengthen relationship. 16
  • 17.
    Critique(Weakness of thepaper) 1. The title of the research was too much 2. It is difficult to judge whether the authors of the research had appropriate academic qualification 3. No recommendation were mentioned in the article. 4. The impact factors of the authors was not a good because only 3 paper cited this paper for the last 7 years. 5. This paper major biased to research project or project but it is not pure scientific research. 17
  • 18.
    Conclusion Generally, this papergood recommender system for social networking friendship and community. The result and methodology based on such forecasting they could found the most popular community among others which would got maximum membership in the future and in which the users are more active. 18
  • 19.