PCOSense AI – An AI-Powered Wellness and PCOS Risk Assessment System

PCOSense AI is an AI-powered wellness and PCOS risk assessment application designed to help users understand and track their wellness information in one place. The application combines Machine Learning, health data tracking, menstrual cycle tracking, personalized insights, and reports through an easy-to-use mobile interface.

The system allows users to create an account and maintain a basic health profile containing information such as age, height, and weight. The application automatically calculates BMI from the available profile information.

The main feature of PCOSense AI is its Machine Learning-based PCOS risk assessment. Users provide selected health and symptom-related information such as menstrual cycle regularity, cycle length, weight gain, hair growth, skin darkening, hair loss, and pimples. The application sends these details to the Django backend, where a trained Logistic Regression model processes the information and generates an estimated PCOS risk probability and risk category.

The system does not diagnose PCOS. Instead, it provides an ML-based risk assessment intended for educational and academic purposes. The application clearly communicates that the results are not a medical diagnosis or medical advice.

PCOSense AI also provides a wellness tracking system where users can record daily information such as sleep hours, water intake, exercise duration, mood, and stress level. This information helps users maintain a history of their wellness activities and understand their personal tracking patterns.

The application includes menstrual cycle tracking, allowing users to record cycle start dates, end dates, cycle length, flow information, symptoms, and notes. Users can review their previous cycle records and manage their tracking history.

Based on the user's assessment, wellness records, and cycle data, the system generates simple rule-based insights and recommendations. These insights can highlight tracked patterns such as low exercise, insufficient water intake, high stress, sleep patterns, or difficult moods. The system avoids making medical diagnoses or unsupported medical claims.

A centralized dashboard brings important information together, including the latest PCOS risk assessment, assessment history, recent wellness records, wellness trends, menstrual cycle information, and generated insights. This allows users to view their tracked information without navigating through multiple screens.

The application also provides a report-generation feature that creates a wellness and risk assessment report containing relevant profile, assessment, wellness, cycle, and insight information. This report can be viewed or shared from the mobile application.

The backend is developed using Python, Django, and Django REST Framework. It provides REST APIs for authentication, user profiles, PCOS risk assessment, assessment history, wellness tracking, menstrual cycle tracking, insights, dashboards, and reports. JWT-based authentication is used to protect user-specific API access.

The machine learning component uses a Logistic Regression model trained on a publicly available PCOS dataset. The trained model is stored separately and loaded by the Django backend when an assessment is requested. The backend validates the input data before passing it to the model and returns the prediction through a REST API.

The mobile application is developed using Flutter and communicates with the Django REST API using Dio. Provider is used for application state management, while Flutter's PDF and printing packages are used for report generation and handling.

PCOSense AI is designed as an academic and demonstration project that combines Artificial Intelligence, Machine Learning, mobile application development, REST APIs, authentication, data tracking, and reporting into a single application. It demonstrates how machine learning can be integrated with a practical wellness tracking system while keeping the prediction clearly separated from medical diagnosis.