Department of computer science and Engineering

A MAJOR PROJECT on

Detection of Child Predators and Cyber Harassers on Social Media

BACHELOR OF TECHNOLOGY
IN
COMPUTER SCIENCE AND ENGINEERING

BY

I CHARAN (20E41A0569)
M TARUN (20E41A0589)
K DHEERAJ (20E41A0597)
M MADHU (20E41A0594)

Under the Guidance of

P ANUPAMA, ASSISTANT PROFESSOR
ACADEMIC BATCH : 2020-2024
CONTENTS

• Abstract
• Existing System
• Proposed System
• Literature Review
• System Requirements
 Abstract

• Professional psychologists must comprehend the risks of cyberbullying and how to shield
children from it.

• The possibility for online sexual conduct is one of the most dangerous aspects of the
internet, despite the fact that it does have some pleasant aspects.

• The major goal of our project is to identify the social media accounts used by predators and
transmit a predator record to the cyber cell administrator (Wolak, 2000; Mitchell, Finkelhor,
& Wolak, 2001).

• This study report provides an overview of our most recent efforts to create a system. As a
result, with an upgraded system, the supervisor can keep taking action after providing any
report to the sexual assault victim
Existing System
• Games, audio chat, and several online entertainment forums all use a variety of game
hunting elements. There is a child predator system that detects online sexual abuse while
children are playing games or utilising online voice chat, protecting the child from
further abuse or sexual exploitation by the abuser since the system is only activated
when children are acting or playing online. any voice conversation.

• Some social media platforms have started integrating basic artificial intelligence (AI)
and machine learning algorithms to assist in content moderation. These systems can help
identify patterns of abusive behavior and prioritize content for human review. However,
these AI systems may still struggle to accurately assess context and intent, leading to
false positives or missed violations.

• Users are encouraged to report suspicious or harmful behavior they encounter on social
media platforms. Reporting mechanisms typically allow users to flag specific posts,
messages, or profiles for review by platform moderators.
 Proposed System

• To distinguish the image from the text in our system, we will only utilise one image
algorithm. Because the Vector Support Machine (SVM) is a learnt model of machine that
uses splitting methods for two-team separation problems, it will offer greater accuracy in
comparison to the current system.

• We will create a train model using regular and unusual phrases and sentences using an
algorithm, and this train model will analyse incoming user postings to determine whether
they are typical or contain abusive content

• By integrating our proposed system into social media platforms, we aim to enhance user
safety and create a more secure online environment for everyone.
 Literature Survey
YEAR TITLE METHODS
This study focused on early methods of detecting
predatory behavior on social media platforms,
"Detecting Predatory Behavior primarily relying on keyword filtering and manual
2010 - 2014 moderation techniques.
on Social Media"

"Natural Language Processing Using NLP techniques, this study analyzed linguistic
2015 - 2018 for Identifying Predatory cues and semantic features in social media content
Language" to identify potentially predatory language patterns.

Focusing on ethical considerations, this study

examined the implications of using AI for content
"Ethical Considerations in AI- moderation on social media platforms, emphasizing
2019 - 2021 transparency and fairness.
based Content Moderation"
 System Requirements

HARDWARE REQUIRMENTS :

• System : i3 or above.
• Ram : 4 GB.
• Hard Disk : 40 GB

SOFTWARE REQUIRMENTS :

• Operating system : Windows8 or Above.

• Coding Language : python
THANK YOU