← Back to projects

AI Diabetes Prediction

I built this during our second-year AIML module as a first try at a full machine learning pipeline. I used a Random Forest model and got around 90% accuracy, but that only came after a lot of data cleaning, preprocessing, and tuning with GridSearchCV. Most of the work was actually on getting the dataset into a usable state before training.

PythonMachine LearningScikit-learnRandom Forest
View source on GitHub →
Google Colab run-through from dataset upload to live predictions.