Classification of Cardiac Electrocardiography with MuyGPs

(In collaboration with Lawrence Livermore National Lab)

• Data exploration and visualization of the ECG datasets.

• Explored different embedding/normalization/feature extraction techniques.

• Applied MuyGPs (a scalable approximate Gaussian process model) and other machine learning models for classification of the ECG datasets

• Explored uncertainty quantifications on the MuyGPs model to demonstrate improved confidence in the models.

See the results of our findings in the arXiv paper: "Enhancing Electrocardiography Data Classification Confidence: A Robust Gaussian Process Approach (MuyGPs)",PDF 

Nuclear Localization Signal (NLS) Prediction

(Erdős Data Science Boot Camp Spring 2024 )

Team Members: Scott Auerbach, Ukamaka Nnyaba, Ming Zhang, Yingyi Guo, Hemaa Selvakumar, and Cisil Karaguzel