With the widespread use of online social networks and their popularity, social networking platforms have provided us with innumerable chances not previously available, and their benefits are evident. Regardless of blessings, people can be embarrassed, ridiculed, tormented, and compelled to utilize the resource of anonymous users, strangers, or peers. In this study, we offered a cyberbullying detection method to build abilities from Twitter content material by employing a pointwise mutual facts technique. We superior a supervised tool learning solutions for cyberbullying detection and multi-beauty classification of its severity in Twitter based on the one's talents. We achieved Embedding, Sentiment, and Lexicon characteristics, as well as PMI-semantic orientation, throughout the test. The extraction of features was conducted utilizing