Today, online communication is accessible to every person in the world, most popularly, in the form of textual messaging. With this advantage comes some challenges, online abuse and harrasement has prevailed in such communication. People often get hate speech, comments and face abuse in the digital world. This problem can be easily solved using Machine Learning and Natural Language Processing. So, for our semester project, we decided to develop an offensive language detection model integrated in a chat app named "Decent Chat." We have use Logistic Regression for binary classification of our messages, offensive and non-offensive. We have also made a reporting mechanism to report such offensive language detected using SMTP to the concerned authority. Here chat messages is our feature and offensiveness is our target.
For training our model accurately, we have taken datasets from various sources of Kaggle. The data was further extracted on the excel sheet and converted into a csv file.
pandas scikit-learn numpy flask_cors flask nltk contractions
cd Api/
python -m venv myenv
source myenv/bin/activate
pip install numpy flask_cors scikit-learn nltk contractions flask
python api.py
npm i nodemon nodemailer
npm start
npm i react-scripts
npm start
Ankit Kumar Karn
Kshitiz Khanal
Royas Shakya
Aditya Shrestha