Time Series extracted from the original dataset representing average signal in 30 minutes interval for subjects in the CMI dataset.

In this case, the Severity Impairment Index (sii) has been binarized (and consequently simplified for tasks involving time series analysis) between non problematic (sii = 0.0) or problematic (sii in [1,2,3]) index.

The file is a gzipped version of a python list where each element is a data frame representing a time series, where on the rows you find the time steps (0-200) and on the columns you find the signal source, the user id and the severity impairment index.

For opening it use


import gzip 
import pickle 

with gzip.open("CMI_timeseries_dataset.pkl.gz", "rb") as f:
	 CMI_timeseries_dataset = pickle.load(f)


Access to time series data frame normally as, for example, CMI_timeseries_dataset[0] (corresponds to first time series)


If you do not have gzip or pickle, pip install them first



