A Profound Method for Three-Tier Toxic Word Classification using LSTM-RNN

Text classification is the activity of labeling natural language texts with relevant categories from a predefined set. In the research work, The words are categorized from a sentence according to their toxicity and according to modification will be done. As Online platforms such as social media provide opportunities to users to express their thoughts and their opinions on various topics and incidents. Due to the different opinion of every individual, an explicit language in some cases which hurt the readers. So, to solve a novel method is proposed by building a model wherein toxic words are categorized into three classes in keeping with their toxicity. In the primary class, words that are least toxic are indicated through crimson traces to the person, withinside the secondary class words are extra poisonous than the primary one, and again and again. The use of those phrases through individual, that words will be eliminated automatically and withinside the remaining or third class, words are the maximum toxic in assessment than that of the first and second class, through the use of those phrases’ person account could be deleted/blocked. As a punishment, the account will be blocked for a specific time period and the length could be improved as in increasing of the occurrences of those words.