Disease Prediction System(DPS)
Prevention Starts with Prediction
Created on 7th September 2025
•
Disease Prediction System(DPS)
Prevention Starts with Prediction
The problem Disease Prediction System(DPS) solves
The Disease Prediction System can be used to:
🧪 Early Diabetes Screening – Quickly check whether a person is at risk of diabetes using basic health details.
⚡ Fast Health Assessment – Enter simple inputs like glucose, BMI, and age to get an instant prediction.
🏥 Support for Doctors & Clinics – Acts as a decision support tool for doctors, helping them screen patients faster before recommending detailed tests.
👨👩👧 Self-Check for Individuals – Anyone can use it from home to get a preliminary idea of their health condition.
🌐 Accessible Anywhere – Runs online on Streamlit Cloud, so no installation or setup is required.
🔑 How It Makes Tasks Easier & Safer
✅ Saves Time – No need to wait for multiple medical tests just for a first-level check.
✅ Increases Awareness – Encourages people to monitor their health regularly.
✅ Assists Healthcare – Doctors can quickly identify high-risk patients.
✅ Accessible & Safe – Works on any device with internet, making health predictions more inclusive.
Challenges I ran into
While building this project, one major hurdle we faced was:
⚡ Mismatch in Features During Prediction
Initially, when we tried to predict the output for a new patient, the app crashed with an error:
ValueError: X has 7 features, but StandardScaler is expecting 8 features as input
This happened because the model was trained on 8 features (Pregnancies, Glucose, Blood Pressure, Skin Thickness, Insulin, BMI, Diabetes Pedigree Function, Age), but during prediction we accidentally passed fewer inputs.
✅ How We Solved It
We carefully checked the dataset columns and matched them with the form inputs.
Added all 8 inputs in the form (Pregnancies, Glucose, BP, Skin Thickness, Insulin, BMI, DPF, Age).
After fixing this mismatch, the model worked correctly and predictions were generated smoothly.
🎯 Learning
Always ensure that the input format for prediction matches exactly with the training data.
Debugging step by step helped us identify where the mismatch occurred.
This taught us the importance of data consistency in ML pipelines.