This document proposes using a clustering algorithm to detect communities in large social media networks. It discusses: 1) Existing community detection algorithms that struggle with social media networks due to their large size and high clustering. 2) A proposed system using K-means clustering and a genetic algorithm to cluster social media data and identify groups of users based on their activities. 3) The advantages of the proposed system in categorizing groups of people, identifying discussion groups, and targeting audiences.

1. SOCIAL MEDIA COMMUNITY USING
OPTIMIZED CLUSTERING ALGORITHM
M.GOMATHI
M.Tech., (Information Technology)
PHASE I

2. • A framework for the novel task of detecting
communities by clustering messages from large
streams of social data. Our framework uses K-Means
clustering algorithm along with Genetic algorithm
and Optimized Cluster Distance (OCD) method to
cluster data.
Abstract

3. Introduction
Social media serves as a ubiquitous public platform which
remains accessible to users as a multiple group of internet
applications. Within the applications, the user creates individual
and unique expression for data exchange.
The study of this data by industry specialists seeking new and
inventive methods to collect data for analysis remains important
to the future of social media.

4. Existing System
The problem of community detection has been widely studied
within the context of large-scale.
Community detection algorithms attempt to identify groups of
vertices more densely connected to one another than to the
rest of the network. Social networks extracted from social
media however present unique challenges due to their size and
high clustering coefficients.

5. • Inability to account for high social dimension and
return results that are precise enough to be useful
in OEC (Online Extremist Community) detection.
DISADVANTAGES FOR EXISTING SYSTEM

6. Proposed System
Clustering structures the data into a
collection of objects that are similar or
dissimilar and is considered an unsupervised
learning. The application of our method is
mainly on finding user groups based on
activities and attitude features as suggested in
the authority model..

7.  This system helps to categorize group of
people.
 This system helps to identify group of people
participated in discussion.
 This system helps to approach targeted
crowd.
 We had used an effective algorithm which
will provide accurate result.
ADVANTAGES OF PROPOSED SYSTEM

8. HARDWARE REQUIRMENTS
SOFTWARE REQUIRMENTS
System : Intel Core i3 2.2 GHz
Hard Disk : 40 GB
Monitor : 14 inch
Mouse : Zebronics.
RAM : 256 MB
O/S : Windows 7 Professional 32bit
Language : Java
Data Base : MySQL
IDE : Net Beans IDE 8.2

9. DFD

10. Module Description
 Building the Network
 Degree Distributions
 Community Structure

11. Building the Network
Generate a social network and to improve the quality and
the representatively of the resulting graph, we filter some of
the comments according to the following four criteria:
The post
Anonymous comments were also discarded
We discard very low quality comments with score
Degree Distributions
The analysis of this degree distribution describes the level of
interaction between users and provides a robust indicator
about the grade of heterogeneity in the network.

12. Community Structure
K-Means algorithm is used to cluster data
and detects communities by clustering
messages from large stream of social data.
This module helps to identify group of people
involved in discussion.

13. SCREENSHOTS – ADMIN LOGIN PAGE

14. SCREENSHOTS – GROUP CREATION

15. SCREENSHOTS – FRIEND REQUEST

16. SCREENSHOTS – PIE CHART for communities

17. APPLICATIONS
• Retail companies market via social networks to is
cover what consumers think about branding,
customer relationship management, and other
strategies including risk prevention.
• Polling within groups and early detection of
emerging events
• Track health issues

18. CONCLUSION
This project proposes a novel method to analyze social
media data. In this method ,the method used to fetch
data is face pager .The fetched data will be export into
csv files.
By separating the post comments according to the
activities and find the result and make the
communities based on the result of post comments.

19. REFERENCE
• A Frame Work for Communities Detection in Social Media
Networks by K.Gowthami M.Tech, Dept. of C.S.E., Adikavi
Nannaya University College of Engg, Rajahmundry, India
• Online Extremist Community Detection, Analysis, and
Intervention by Lieutenant Colonel Matthew Curran Benigni
June 2016
• Social Network Clustering by Narges Azizifard Department of
Computer Science and Engineering of Islamic Azad University
of Qazvin Branch, Qazvin, Iran

20. Thank You…