Machine Learning Disease Prediction System

What Inspired Us

We were inspired by the healthcare challenges in many low- and middle-income countries, including our own. People in remote or underserved areas often delay seeking care because they can't reach a doctor or aren’t sure if their symptoms are serious. We wanted to create a tool that brings basic health support closer to these communities.

What We Learned

We learned how to collect, clean, and use real-world health data to train machine learning models. We also explored the importance of user-friendly design for health tools. Most importantly, we saw how technology can be used to solve real public health problems.

How We Built It

We used a dataset from Kaggle with over 4,900 records and 134 symptoms. We trained a Random Forest Classifier in Python to predict diseases based on user-inputted symptoms. For deployment, we used Flask to build a web app where users can enter their symptoms and instantly see a predicted disease. We also used libraries like Pandas, NumPy, scikit-learn, and Joblib.

Challenges

• First challenge Managing our university coursework while meeting regularly to build and test the project was a challenge, one of our teammate had exams as well but we learned a lot. Grateful for the experience

• we had different schedules and skill levels. Learning how to divide tasks and support each other was part of the journey.

• We frequently ran into small bugs in the code or errors during model training and didn’t always know how to fix them right away

• Also, making sure the model worked well even with fewer symptoms entered.

Built With

• flask
• html/css-(for-basic-frontend)-libraries-&-frameworks:-scikit-learn-?-for-building-the-machine-learning-model-pandas
• joblib
• numpy
• pandas
• python
• scikit-learn
• vscode