suparnasingh27
Jun 24 min read
What Kind of Aggregation Should Be Used When Working with the Sepsis Dataset?
The Sepsis dataset (https://physionet.org/content/challenge-2019/1.0.0/) contains data for 40,336 patients, with hourly ICU readings of vital signs, lab values and clinical observations. When working on identifying KPIs, gathering insights and establishing correlations between the biomarkers, I did not pay much attention to the types of aggregations that should have been used i.e., average, minimum and maximum. At the beginning of the analysis, I had assumed that calculating






