Social Media Cyber bullying Detection using Machine Learning Preprocessing Preprocessing involves cleaning up the data by removing noise and extraneous text. 1.1 Feature Extraction The features extraction stage is the second step. The textual input is changed in this step into a format that may be used to feed machine learning algorithms. In this step, TFIDF and sentiment analysis are applied. 1.2 Detection and Prevention measures for Cyberbullying and Online Grooming In this system, individuals can post messages and photographs on a social media network similar to Facebook. The posting of adult photos, inappropriate comments, and other material is prohibited by this system. The User Messages are categorised using the Bad Words Dataset and the Sensitive Words Dataset in this case, and the user's bad count is entered into the database. Any user who submits too many offensive or pornographic photographs will immediately be banned from this social networking site. 1.3 Adult Image Detection Any user who submits an adult image will receive a warning. Before their account is blocked, each user has a limited number of opportunities. If the user is discovered to be submitting this type of stuff, there is a clause for the database item that determines whether or not the account may be kept open. 1.4 Irrelevant Posts Detection Algorithm This algorithm extracts keywords from text files containing different categories of data and user messages are scanned to find whether they contain those keywords to classify messages into the Crime/Worst/Riots category and to find the sentiment of a message. 1.5 NLP Algorithm A approximate idea of a user's social standing can be obtained by analysing comments, the percentage of positive or negative responses, and sentiment analysis using different text mining modules. User posts are processed into tree structures, and the words within those tree structures are subsequently analysed to determine the posts' sentiments.

#