A Review of Deep Learning Methods for Irregularly Sampled Medical Time Series Data

Chenxi Sun, Shenda Hong, Moxian Song, and Hongyan Li. https://arxiv.org/abs/2010.12493

Irregularly Sampled Medical Time Series Datasets

Name	Folder
Sepsis early diagnosis	/data/sepsis_data
COVID-19 mortality classification	/data/covid19_data
MIMIC-III in-hospital mortality classification	/data/mimiciii_inhospital_mortality_data
ICU mortality data	/data/icu_mortality_data
Synthetic patient data	/data/emrbots_data

Other Time Series Datasets

(Can be regular or irregular data)

Name	Folder	Link
SILSO solor dataset	/other_data/solor_data	https://lasp.colorado.edu/lisird/data/international_sunspot_number
USHCN weather dataset	/other_data/usa_weather_data	https://www.ncei.noaa.gov/products/land-based-station/us-historical-climatology-network
Synthetic medical records	/data/emrbots_data	http://www.emrbots.org
Chaotic system datasets	/other_data/chaotic_system_data
UCR comprehensive datasets	/	https://www.cs.ucr.edu/~eamonn/time_series_data_2018
UEA comprehensive datasets	/	http://www.timeseriesclassification.com/index.php
Forecasting comprehensive datasets	/	https://github.com/thuml/Time-Series-Library
MIT-BIH arrhythmia ECG dataset	/	https://physionet.org/content/mitdb/1.0.0/
MIT-BIH ST change ECG dataset	/	https://physionet.org/content/stdb/1.0.0/
European ST-T ECG dataset	/	https://physionet.org/content/edb/1.0.0/
Sudden cardiac death holter ECG dataset	/	https://physionet.org/content/sddb/1.0.0/

Category of related Works

Conclusion of related Works

Sturcture of EHRs

Missing rates of the real-world EHRs datasets

Sturctures of RNN-based methods