Cyber bullying Detecation

Project Title

Cyber bullying Detecation

Project Area of Specialization Computer ScienceProject Summary

Cyberbullying is the sending of abusive messages over electronic communication. Sharing someone's private information on social media to embarrassed or harass someone is also include in cyberbullying. It can be through Social media platforms and instant messages. It can lead users to mental problems such as stress, and depression. Generally, the generation that commonly involves with cyberbully is the young generation. People didn't know whether their actions like comment, post status, share the post and also personal message become a threat toward others. Unfortunately, most people have depression on using social media due to cyberbully. There are many solutions available but users cannot easily accessible for the user. So we are making a system that is accessible for the user and userfriendly for cyberbullying detection by a machine learning approach that helps the user to know about cyberbullied words and the user can save himself from repetitive abusive curl languages or hate speeches. The system detects text and detects their category such as racist, sexual, physical mean or others, by using a machine learning approach. The importance of the system is to analyses the status of the people’s whether the people’s is being cyberbullies or cyberbullied or in the middle between being cyberbullies and cyberbullied Through this identification system, the peoples can know their category of Cyberbully and that they can gain with the ideas and also consultation to beat the Cyber bully?s problem.

Project Objectives

The goal of this project is to the prediction of cyberbullying-related text on social media. Manual monitoring for cyberbullying is very difficult and time-wasting. Automatic detection of cyberbullying would enable moderation and it's easy to respond quickly when it's need. The main aim of this project is that it presents a system to that predict detection of cyberbullying, including different types of cyberbullying. It also works as when user copy past massages on website is detect that the massages is cyberbullied or not if it is cyberbullied its check in which category it lies.

Project Implementation Method Implementation

To run a system properly, the back-end coding, front end interface, and their connectivity must all be implemented. This section covers all the back-end coding required to build that project. It also comprises the system's graphical user interface (GUI). At this point, all the features have been merged to provide a meaningful image. Machine learning web based application in the field of Cyber bullying Detection could be most appropriate technique categories and recognize Cyber bulled text, which could be highly useful in the daily life.

5.1 Algorithm

The algorithms that are integrated are discussed here.

5.1.1 Dataset

All the data is in a text from, the dataset consist of different cyberbully text.

5.1.2 Preprocessing

To prepare the text data for the model building we perform text preprocessing. It is the very first step of NLP. Some of the preprocessing steps are:

• Removing punctuations like. , ! $( ) * % @
• Removing URLs
• Removing Stop words
• Lower casing
• Tokenization
• Stemming
• Lemmatization

5.1.3 Prediction

5.1.3.1 NLP 

The NLP is implemented in that manner as follows:

5.1.3.1.1 Pre-processing

All Pre-processing techniques are used.

5.1.3.1.2 Data Validation  

We used Data validation process.

5.1.3.1.3 Classification.

We used Random forest Algorithm for Classification.

5.1.3.2 Implementation  

1. Input Dataset

Pickleibary of python is used to load dataset in algorithm.

1. Random Forest Model

Random Forest Model are used.

1. Training

The Data are divided into testing and training. We used 1100 entries  for training the

data. All factors of training phase are achieved like updating of weights, loss error function

and optimization of algorithm. This process continued. After training, the model we save

the model with pyTorch (.PTH) file extension.

1. Testing

We used 500 entries for testing from each cyberbullying classes. From pre-trained model, test data are evaluated, and classification is done.

1. Output

When all steps are done, accuracies of all individual classes are displayed on GUI screen.

Benefits of the Project

Cyberbullying as the name implies is the use of cyberspace as a mechanism to bully others known or unknown to the bully. Cyberbullying has caused significant issues for those involved ranging from extreme displays of anger to suicide attempts. The social media network is a great platform for us where we connect with each other daily. Cyberbullying on a social media platform is a global issue in know days because of its large number of active users. The cyber bullying on social network is growing rapidly every day. According to a recent report that cyberbullying constitutes a growing problem among youngsters. Successful prevention depends on the adequate detection of potentially harmful messages and the information overload on the Web requires intelligent systems to identify potential risks automatically. So, in this project we focus on to make a web based cyberbullying detection system which detection Text and help people to stop cyberbullying.

The problem that usually arises when peoples did not know whether their actions such as comment, post status, share the post and also personal message on social media become a threat toward others. Many social media users don't know about cyberbullying and how to overcome it. Most of the peoples that been experienced cyberbully feels embarrassed to share with others about being a cyberbully and they tend not to give attention about it. Therefore, based on all problem statements it is important to develop a web-based system to help peoples overcome cyberbullying.

The main objective of this system is to solve the problem that social media users are facing nowadays. • To develop an efficient system for cyberbully identification from social media. • System is to detect categories of people in which category the user belongs. i.e. cyberbullies or cyberbullied or in-between them. • Detect abusive text messages and comments from social media. • Cyberbullying can be very damaging to teenagers. It can lead to mental problems and even suicide. So, we developed a system to overcome these issues.

Technical Details of Final Deliverable

DELIVERABLES FOR Cyberbullying detection

• Requirement specification (document)
• User interface

We make user interface by using Django framework. we use Html ,CSS, bootstrap for this website. Frist, we make a web interface and make a singin in and singup interface for user user first make account after that they get a interface to of textbox they can copy paste their text in that text box and analyze the text. Then the machine learning algorithms detect that text bulling or not and also categorize text into four category.

Harassments, islamophobia etc

• Backend development

For database we use sql all data store in this database.

• Model training

In model training we used dataset and train the model

• Set up of Test system

Then we test the system by checking different comments on social media the accuracy is to good.

• Set up of Live system

Then we live up the system.

• User training

In user training we train user how to used the system.

Final Deliverable of the Project Software SystemCore Industry TelecommunicationOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Quality Education, Gender Equality, Life on LandRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	40578
Hikvision SSD C100 Series 512GB 2.5	Equipment	2	9990	19980
Kingston HyperX FURY 4GB 288-Pin DDR4 24	Equipment	2	6999	13998
USB	Miscellaneous	1	1600	1600
others (printing ,poster,Docments)	Miscellaneous	2	2500	5000