-
Home page
-
Assessment details for analysis
-
Assessment Result
-
Cara-pcos Support
Inspiration
Polycystic Ovary Syndrome (PCOS) affects 1 in 10 women worldwide, yet 70% remain undiagnosed. Watching friends struggle with unexplained symptoms for years before getting diagnosed inspired us to create FemCare. We realized early detection could prevent serious complications like diabetes, heart disease, and infertility. Traditional diagnosis often takes 2-3 years and multiple doctor visits – we wanted to change that equation.
What it does
FemCare is a comprehensive platform that: 1.AI-Powered Risk Assessment: Analyzes 15+ health parameters using machine learning to calculate PCOS risk probability 2.Personalized Recommendations: Provides risk categorization with specific lifestyle and medical advice 3.24/7 Chat Support: "Cara" – our PCOS-specialized chatbot offers immediate support on symptoms, diet, exercise, and mental health 4.Educational Resources: Statistics, symptoms information, and management strategies in an accessible format
How we built it
Frontend: HTML5, CSS3, JavaScript with responsive design Backend: Flask (Python) web framework ML Model: Scikit-learn Random Forest Classifier Chatbot: Rule-based NLP system with regex pattern matching Deployment: Local server with Flask development server
Challenges we ran into
1.Imbalanced Dataset: PCOS vs non-PCOS cases were uneven (70:30) Solution: Used SMOTE oversampling and class weighting 2.Chatbot Context Understanding: Users ask questions in diverse ways Solution: Implemented regex patterns + keyword combinations 3.Feature Correlation: Many symptoms correlate (e.g., weight gain and insulin resistance) Solution: Used feature importance scores and VIF analysis 4.Medical Accuracy vs Simplicity: Balancing medical precision with user-friendly explanations Solution: Created tiered responses with optional detailed information
Accomplishments that we're proud of
1.High Prediction Accuracy: 92% accuracy validated with 5-fold cross-validation 2.Zero-Dependency Chatbot: Works completely offline with no API costs 3.Complete User Journey: From assessment to support in under 3 minutes 4.Accessibility Focus: Designed for users with varying tech literacy 5.Privacy First: No personal data storage, all processing local
What we learned
1.Medical AI Ethics: Importance of clear disclaimers and encouraging professional consultation 2.Feature Engineering: How seemingly minor symptoms (like skin darkening) significantly impact predictions 3.Women's Health Tech Gap: Limited accessible tools for common conditions like PCOS 4.Regex Power: Surprisingly effective for domain-specific chatbots when well-designed 5.UI Psychology: Color schemes and wording significantly impact medical tool adoption
What's next for FemCare
1.Mobile App Development: iOS/Android versions 2.Symptom Tracker: Daily logging with trend analysis 3.Doctor Connect: Platform to share results with healthcare providers 5.Multilingual Support: Starting with Spanish and Hindi 6.Personalized Plans: AI-generated diet/exercise regimens
Log in or sign up for Devpost to join the conversation.