Suicide Detection From User Input

Link to run and test

In this project we are using different machine learning and deep learning model to detect if a message is suicidal or not

---

Project Structure

• app/ : Api folder contains the api using FastAPI
• dataset/ : You can put your dataset here. link
• inference.py : Script to predict and classify using pre-trained model
• model.joblib : Pre-Trained model using naive bayes classification
• requirements.txt : Requirements file
• train_nb.py : Script to train model

---

Train

To train the dataset based on your own specifications you can go through the following steps:

1. Download dataset and place it in the dataset folder
2. Create your virtual environment using: python venv
3. Install requirements using the command: pip install requirements.txt
4. Change parameters such as max_features in text_processing.py
5. Run train_nb.py and save your model

---

Predict

To predict you can use pre-trained model and run inference.py

---

Future Works

• Train model using deep learning methods
• Deploy api
• Enhance translation and add more languages