Cyber Bullying Detection Using Artificial Intelligence

Abstract

Social media has given us a lot of opportunities and benefits. In spite of all the benefits, people are still getting bullied by anonymous users. When someone bullies another person using technology, they may send or post offensive or hateful messages about them. This is known as cyberbullying. With the extensive use of social media users, cyber bullying has also increased. Hence, it’s very important to detect cyber bullying on online platforms. This enables large-scale social media monitoring. Here we have implemented this project to identify the hate speech or offensive speech from twitter. Main aim of our project is to detect cyber bullying in tweets using Machine learning and deep learning classification algorithms like Logistic Regression, Naive Bayes, LSTM (Long Short-Term Memory) and CNN (Convolution Neural Network). NLTK (Natural Language toolkit) is also used to preprocess text data. Then compare all the algorithms used, check the accuracy of each model and then choose the best model for detecting cyber bullying in tweets.  The motivation behind this project is to protect our society from cyber bullying and also to prevent youngsters and teenagers from getting bullied, committing suicide because of bullying and reducing crime in cyberspace.