ML Model for Cyber Bullying Detection in Texts
ML model that can effectively detect instances of cyber bullying in text-based communications, such as social media posts, comments, chat messages, and online forum.
Created on 25th June 2023
•
ML Model for Cyber Bullying Detection in Texts
ML model that can effectively detect instances of cyber bullying in text-based communications, such as social media posts, comments, chat messages, and online forum.
The problem ML Model for Cyber Bullying Detection in Texts solves
The ML model for cyber bullying detection in texts solves the problem of identifying instances of cyber bullying in online communications. Cyber bullying refers to the use of electronic communication platforms, such as social media, messaging apps, and online forums, to harass, intimidate, or harm individuals. It can take various forms, including sending offensive messages, spreading rumors, making derogatory comments, or engaging in personal attacks.
The problem this ML model addresses is significant because cyber bullying can have severe psychological and emotional impacts on the victims. It can lead to decreased self-esteem, increased stress and anxiety, depression, and even suicidal ideation. Detecting and mitigating cyber bullying is crucial to protect individuals' well-being and create safer online environments.
Challenges we ran into
We initially planned to use a Naive bayes model using tokenization and dictionaries but due to a very dirty data the acuuracy was very low. The data set we took was of tweets so many special characters caused us some trouble in the preprocessing and cleaning stages. We tried to use the lazypredict model to find the optimized model to build for our project but the data failed us there as well. We had to test over 10 models to find one that would work efficiently with real world applications.
Tracks Applied (1)
Replit
Replit
Technologies used