LLNL_DataScience

2024 Lawrence Livermore National Laboratory Data Science Challenge

Challenge Details:

• Task 1 : Heartbeat Classification

  Get familiar working with ECG data by using the ECG Heartbeat Categorization Dataset to perform binary classification for healthy heartbeat vs. irregular heartbeat

• Task 2 : Irregular Heartbeat Classification

  Diagnosing an irregular heartbeat by using the ECG Heartbeat Categorization Dataset to perform multiclass classification to diagnose the irregular heartbeats.

• Task 3 : Activation Map Reconstruction from ECG

  Sequence-to-vector prediction using the Dataset of Simulated Intracardiac Transmembrane Voltage Recordings and ECG Signals to perform activation map reconstruction (i.e. transform a sequence of length 12x500 to 75x1 using a neural network)

• Task 4 : Transmembrane Potential Reconstruction from ECG

  Sequence-to-sequence prediction using the Dataset of Simulated Intracardiac Transmembrane Voltage Recordings and ECG Signals to perform transmembrane potential reconstruction (i.e. transform a sequence of length 12x500 to 75x500 using a neural network)
• Learn more: Original LLNL Data Science Introduction Repo

Datasets:

• Task 1 & 2
• Task 3 & 4

About the files:

• Personal notebook
  • Walks through the learning process of the tasks (1-3)
  • Discusses machine learning methods/techniques
  • Runs through the training/testing models
  • Displays diagrams of datasets
• Personal test notebook
  • Secondary notebook for processing data
  • Trains/Processes task 1-3 data while running personal notebook
  • More diagrams for data
• Task 3 and 4 test notebook
  • Mainly task 3 data manipulation
  • Development of a model for task 3 classification
  • Walks through the individual layers of the model
• Task 3 and 4 final model
  • Finalized 1D Squeezenet model for heart ECG activation/lead prediction