work
Machine Learning Intern at HealthcareSystems R&A
Project type
Machine Learning Intern at HealthcareSystems R&A
Year
2023
Project overview
During my time as a Machine Learning Intern at Healthcare Systems R&A in Montreal, I had the opportunity to dive deep into the world of healthcare analytics. Our team was tasked with developing advanced machine learning models to improve cardiac care, specifically focusing on predicting the need for cardiac catheterization.
No items found.
What I did
I developed a custom bagging neighbor model for cardiac catheterization prediction using healthcare data. To address imbalanced datasets, I implemented SMOTE and SMOTE-ENN, improving balanced accuracy by 15%. I applied PCA and Isomap for dimensionality reduction and feature importance. I also created a PyTorch neural network for predicting adverse heart conditions, achieving 85% accuracy using RayTune for hyperparameter optimization. The project involved iterative development, cross-validation, and Bayesian optimization to ensure reliable predictions.
