AI Hackathon Project: iOS App with Insurance Prediction

Inspiration

Our inspiration came from the growing need for efficient healthcare management, specifically in the area of insurance claim prediction. With many patients struggling to determine appropriate insurance plans, we wanted to build a tool that could help bridge this gap using technology. By creating an easy-to-use iOS app, we aimed to assist patients in predicting the best insurance plan based on their medical history and other user-specific details.

What We Learned

Throughout this project, we learned the importance of data-driven decision-making and how machine learning can simplify complex real-world problems. Working with classifiers like Random Forest and KNN provided insights into the differences in model performance and accuracy. We also gained a deeper understanding of front-end development using React Native, especially in creating user-friendly interfaces for iOS applications.

Project Development

We built the app using React Native for the front-end, ensuring a smooth, responsive design that allowed patients to input their user details easily. For the backend, we used Firebase for real-time data connection, enabling secure and seamless data transfer between users and the prediction model. The app collected relevant information from users and provided an insurance prediction based on their input using machine learning models integrated on the backend.

Challenges Faced

One of the major challenges we encountered was generating a suitable dataset. In the absence of real-world insurance data, we had to create a synthetic dataset that would be realistic enough for model training. Another challenge was identifying the correct classifier for our prediction model. Initially, we experimented with Random Forest, but we found its performance was not optimal for our dataset. After further testing, we switched to the K-Nearest Neighbors (KNN) algorithm, which gave us better accuracy and more reliable predictions.

Built With

• firebase
• matplotlib
• pandas
• python
• react-native
• sci-kit