Kalana-S/AI-Based-Mental-Health-Screening-System

This is a machine learning-powered web application built with Python, Flask, and SHAP, designed to screen for Depression, Anxiety, and Stress using the standardized DASS-42 questionnaire. It provides explainable AI (XAI) outputs, user history, interactive plots, and bilingual (English/Sinhala) support. # 🧠 AI-Based Mental Health Screening System This is a machine learning-powered web application built with **Python, Flask, and SHAP**, designed to screen for **Depression, Anxiety, and Stress** using the standardized **DASS-42** questionnaire. It provides **explainable AI (XAI)** outputs, user history, interactive plots, and **bilingual (English/Sinhala)** support. --- ## 🚀 Features - Predicts **Depression, Anxiety, and Stress** severity levels (Normal / Moderate / Severe). - Built with **Logistic Regression** and SHAP for transparent predictions. - Clean and bilingual UI with **light/dark mode** and responsive design. - Supports **anonymous** mode or **registered accounts** with history tracking. - Interactive **plots** for trends over time using Plotly. - Personalized **tips & recommendations** based on results. - Admin panel for **content management** and **system logs**. --- ## 🩺 Input Parameters - DASS-42 questionnaire responses (42 questions) - User account details (if registered) --- ## 🧰 Technologies Used - **Python** – Core development - **Flask** – Web application framework - **Pandas, NumPy, Scikit-learn** – Data processing & ML model training - **SHAP** – Explainable AI visualizations - **Plotly** – Interactive plots and graphs - **HTML, CSS, Bootstrap** – Frontend UI - **Pickle, JSON** – Model & metadata storage --- ## 📁 Files Included - `app.py` – Flask app entry point - `templates/` – HTML templates for pages (home, login, assessment, results, history, tips, admin) - `static/` – CSS, JavaScript, and images for frontend styling - `pickle files` – Pickled ML models (`depression.pkl`, `anxiety.pkl`, `stress.pkl`) - `scaler.pkl` – Preprocessing scaler for consistent inputs - `dass42_questions.json` – Questionnaire text in English & Sinhala - `requirements.txt` – Project dependencies - `README.md` – Project documentation - `ML_Model.ipynb` – Jupyter notebook for training models --- ## ⚙️ Installation & Setup 1. Clone the repository: ```bash git …