Racism-and-Offensive-Text-Detection

Link to the Google Colab code to the main code: https://colab.research.google.com/drive/1T-9GJBWCgE5QxDgEgSJ6p9YQYcYosTx7?usp=sharing
Link to the Google Colab code to the deployment based code: https://colab.research.google.com/drive/1KHUlQZHDxMm9iEGz2TLFJQt6w9K5AKBD#scrollTo=SaketWGJWZ6D
Here, I've created models which predict whether a given sentence contains some hate or offensive overtone. The logistic regression classifier has done well in both the cases.

We get an accuracy of around 94% in both the cases of hate and offensive content detection.
This model has been deployed using Flask. Other frameworks used include

1. HTML
2. CSS
3. Python

The code for the deployed model is included here as well along with all the other frameworks as well.
However, in the jupyter notebook, the models are pickled and available in this repository. The other methods I have used here (only in the jupyter notebook) include: Naive Bayes, Decision Tree, KNN and Random Forest.

Site functionality

The main page would take in the user input and once we give in an input, click on predict. (Refer page_1.jpg)
Then, based on the input, the output would be provided. (Refer page_2.jpg and page_3.jpg)