Welcome to the Student Depression Detection project—where we dive into the serious issue of mental health among students, armed with numbers, neural networks, and a dash of good intention.
"They say laughter is the best medicine, but sometimes, a little data science can help too."
Project Overview
Mental health is no laughing matter, and for students, the struggle can be real. With academic pressures, late-night study sessions, and caffeine-powered diets, it's no wonder mental well-being often takes a backseat.
This project uses a machine learning model (PyTorch-powered) to predict signs of depression among students based on factors like dietary habits, sleep duration, family history, and study hours. While it's no replacement for professional help, this tool aims to shine a light on potential risks and help identify when someone might need support.
How it works
This project leverages a dataset that examines the connection between mental health and diverse factors, including:
- Demographics: Gender and age.
- Academic Factors: Academic pressure, study hours, and satisfaction levels.
- Lifestyle Choices: Sleep duration, dietary habits, and financial stress.
- Mental Health Indicators: Family history of mental illness, depression status, and suicidal thoughts.
The dataset allows for an exploration of how these variables—such as sleep quality, diet, and academic workload—interact and influence mental well-being. The goal is to identify patterns of mental health risks and to promote preventive strategies for improving outcomes among various demographic groups.
Source : https://www.kaggle.com/datasets/ikynahidwin/depression-student-dataset (Thanks to them!)
Pipeline :
- Preprocessing: Data cleaning, feature encoding, and scaling are applied to ensure compatibility with the model.
- Modeling: A neural network processes the data, learns patterns, and predicts whether a student shows signs of depression.
- Evaluation: Metrics like accuracy, precision, recall, and ROC-AUC are computed, with visualizations (e.g., confusion matrix and ROC curve) providing further insights.
- The result is a system capable of identifying potential mental health risks based on a combination of academic, demographic, and lifestyle factors.
📊 Results
The model performed well, with an impressive ROC-AUC score, but here's a friendly reminder: This model is not a diagnostic tool. Always consult professionals for mental health concerns.
🤝 Contributing
Got ideas for improvement? Found a bug? We welcome contributions! Open an issue or submit a pull request.
This tool is educational and experimental. It should not replace medical advice or professional mental health services. If you or someone you know is struggling, please reach out to a qualified mental health professional.
🧠 Mental Health Matters
Remember, it's okay to not be okay. Take breaks, eat well, sleep enough, and don’t hesitate to ask for help. Oh, and let’s be kind to ourselves—debugging life is harder than debugging code.