P12data/ReadMe.md

PhysioNet 2012 Challenge dataset

Raw data###

• /rawdata/
  • Raw data: set-a, set-b, set-c, Outcomes-a.txt, Outcomes-b.txt, Outcomes-c.txt
  • Data source: https://www.physionet.org/content/challenge-2012/1.0.0/

Data parsing (process_scripts)

Run following scripts in turn:

• /process_scripts/
  • ParseData.py : generate arr_outcomes.npy, ts_params.npy, static_params.npy, and P_list.npy
  • IrregularSampling.py: generate: extended_static_params.npy, PTdict_list.npy
  • Generate_splitID.py: generate phy12_splitX.npy where X range from 1 to 5. Only contains the IDs. Train/val/test = 9:1:1

Note: PTdict_list.npy and arr_outcomes.npy are the most important files.

Processed data ###:

• /processed_data/

  • PTdict_list.npy

  • arr_outcomes.npy

  • ts_params.npy

  • extended_static_params.npy

  • static_params.npy

  • P_list.npy

• /splits/
  • phy12_splitsX.npy where X range from 1 to 5. In splits folder: there are 5 npy files, each contains three array representing idx_train,idx_val,and idx_test.