Social networking platforms have given us incalculable opportunities than ever before, and its benefits are undeniable. Despite benefits, people may be humiliated, insulted, bullied, and harassed by anonymous users, strangers, or peers. Cyberbullying refers to the use of technology to humiliate and slander other people. It takes form of hate messages sent through social media and emails. With the exponential increase of social media users, cyberbullying has been emerged as a form of bullying through electronic messages. We have tried to propose a possible solution for the above problem, our project aims to detect cyberbullying in tweets using ML Classification algorithms like Naïve Bayes, KNN, Decision Tree, Random Forest, Support Vector etc. and also we will apply the NLTK (Natural language toolkit) which consist of bigram, trigram, n-gram and unigram on Naïve Bayes to check its accuracy. Finally, we will compare the results of proposed and baseline features with other machine learning algorithms. Findings of the comparison indicate the significance of the proposed features in cyberbullying detection.
Introduction
I. INTRODUCTION
Cyberbullying can be defined as an aggressive or intentionally carried out harassment by group or individual through digital means repeatedly against a sufferer who is unable to defend themselves. This type of bullying includes threats, abusive or sexual remarks, rumours and hate speech. Cyberbullying is an ethical issue found on internet and the percentage of the victims is also alarming.
A. Cyberbullying on Social Media Sites
The major contributors to cyberbullying are social networking sites. The dynamic nature of these sites helps in the growth of online aggressive behaviour. The anonymous feature of user profiles increases the complexity to identify the bully. Social media is popular due to its connectivity in the form of networks. But this can be harmful when rumours or bullying posts are spread into the network which cannot be easily controlled. Twitter and Facebook can be taken as examples which are popular among various social media sites. According to Facebook users have more than 150 billion connections which gives the idea about how bullying content can be spread within the network in a fraction of time. To manually identify these bullying messages over this huge network is difficult. There should be an automated system where such kinds of things can be detected automatically thereby taking appropriate action. The victims mainly consist of women and teenagers. Intense effect on mental and physical health of the victims in such kind of activities higher’ s the risk of depression leading to suicidal cases. Therefore to control cyberbullying there is need of automatic detection or monitoring systems.
???????B. Detection Models for Cyberbullying
Many works in this field have shown that machine algorithms can be used for predicting and detecting cyberbullying actions. As tremendous data is generated every second algorithms can be trained efficiently. Machine learning classifiers help to classify the content of the texts into non - bullying and bullying classes. After classification of the content the bullying ones can be stopped. This project aims to detect cyberbullying in tweets using ML Classification Algorithms. A training and predicting pipeline is implemented to contrast performance of various popular classification algorithms and determine the best suited model. Extracted features will be applied with Naïve Bayes, KNN, Decision Tree, Random Forest, Support Vector Machine algorithm etc. and also we will apply the NLTK(Natural language toolkit) which consist of bigram, trigram, n-gram and unigram on Naïve Bayes to check its accuracy. We will compare the algorithm and draw the results, results will indicate that our proposed framework provides a feasible solution to detect cyberbullying behavior and its severity in online social networks.
The motivation behind this project is to prevent teenagers from getting depressed or committing suicide due to cyber bullying activities and also to decrease the harassing incidents in cyberspace.
II. PROBLEM FORMULATION
The social media network gives us to great communication platform opportunities they also increase the vulnerability of young people to threatening situations online. Cyberbullying on an social media network is a global phenomenon because of its huge volumes of active users. The trend shows that the cyber bullying on social network is growing rapidly every day. Recent studies 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 model on automatic cyberbullying detection in twitter by modelling posts written by bullies on twitter. We have tried to propose a possible solution for the above problem, our project aims to detect cyberbullying in tweets using ML algorithms like Naïve Bayes, KNN, Decision Tree, Random Forest, Support Vector etc. and also we will apply the NLTK(Natural language toolkit) which consist of bigram, trigram, n-gram and unigram on Naïve Bayes to check its accuracy. We will train the model using different algorithms and compare the results and check whether the model is able to provide the desired results or not that is whether it is able to predict the text is bullied or not.
III. LITERATURE REVIEW
IV. METHODOLOGY
This project will be developed using python, ML and web technology.
A. First we will search and find the dataset from Kaggle or Github and download it to train the model.
B. After downloading we will pre-process the data, clean the data then with the help of naïve bayes, SVM (Support vector machine), KNN and Random Forest we will train the dataset and generate models separately and also we will apply the NLTK(Natural language toolkit) which consist of bigram, trigram, n-gram and unigram on Naïve Bayes to check its accuracy.
C. We will fetch the real time tweets from twitter and then we apply the generated model to these fetched tweets and check if the text is cyberbullying or not, also we will compare different algorithms and check the accuracy of various models and select the best model.
D. For the frontend purpose we will create a platform where there will be different logins for user and admin and the user will be able to post the tweet and the admin will be able to view the tweets and classify whether the tweet is bullied or not.
E. These all-purpose we are using python as backend, for training the model we are using ML algorithms.
The flow chart depicting the flow of our system is-
The dataset we collected for training the model is-
Pre-processing was carried out by first changing all the uppercase words to lowercase. The next step was to remove all the punctuation marks and emojis as they are not required for our purpose. The next important step is to remove the stop words. They are the common words which are not useful in detection like “a”, “on” , “all” These words do not carry important meaning and are usually removed from texts. A text file containing bad words was created. If the tweet has the words present in the bad words then it is labelled as “TRUE” else labelled as “FALSE”. This final dataset contains the tweets and classification which is now used for algorithm implementation.
Our proposed framework provides a feasible solution to detect cyberbullying behaviour and its severity in online social networks like twitter. After completion of the project we will get a system which would successfully detect the offensive text or comments or the cyberbullied statements on Twitter. The system will help the society especially the teenagers and adults for preventing them from further bullying.
V. RESULT DISCUSSIONS
The system is successfully able to predict whether the content is bullied or not. We have achieved the following output-
VI. ACKNOWLEDGEMENT
We as the authors would like to extend a special thanks of vote to the reviewers of this paper for their valuable suggestions to improve this paper. The paper is supported by Acropolis Institute of Technology and Research, Indore (M.P.)
Conclusion
The study reviewed the existing literature for various machine learning algorithms and identified Naïve bayes N-gram gives the best accuracy and also the system is able to identify the bullied and non-bullied statements. The goal of this project is to the automatic detection of cyberbullying-related posts on Twitter. Automatic detection of signals of cyberbullying would enhance moderation and allow them to respond quickly when necessary. However, these posts could just as well indicate that cyberbullying is going on. The main aim of this project is that it presents a system to automatically detect signals of cyberbullying on social media handle Twitter, including different types of cyberbullying, covering posts from bullies, victims and bystanders.
References
[1] Poeter. (2011) Study: A Quarter of Parents Say Their Child Involved in Cyberbullying. pcmag.com. [Online].Available: http://www.pcmag.com/article2/0,2817,2388540,00.asp
[2] J. W. Patchin and S. Hinduja, “Bullies move Beyond the Schoolyard; a Preliminary Look at Cyberbullying,” Youth Violence and Juvenile Justice, vol. 4, no. 2, pp. 148–169,2006
[3] Anti Defamation League. (2011) Glossary of Cyberbullying Terms.adl.org.[Online].Available:http://www.adl.org/education/curriculum connections/cyberbullying /glossary.pdf
[4] N. E. Willard, Cyberbullying and Cyberthreats: Responding to the Challenge of Online Social Aggression, Threats, and Distress. Research Press, 2007.
[5] D. Maher, “Cyberbullying: an Ethnographic Case Study of one Australian Upper Primary School Class,” Youth Studies Australia, vol. 27, no. 4, pp. 50–57, 2008.
[6] https://www.sciencedirect.com/topics/computer-science/deep-neural-network
[7] An Effective Approach for Cyberbullying Detection and avoidance IEEE paper
[8] Approaches to Automated Detection of Cyberbullying: A Survey IEEE paper
[9] Cyberbullying Detection System on Twitter IEEE paper
[10] Methods for Detection of Cyberbullying: A Survey IEEE paper