Skip to content

The objective of this work is to analyze large datasets provided by Tidepool to determine various actionable insights and refined knowledge that can accelerate the pace of clinical translation of artificial pancreas (AP) and advisory decision support systems for daily use.  Data corruptions such as artifacts, noise, missing values and outliers, unmeasured disturbances, and human errors mask the true fidelity of the sensor data collected in free-living conditions.  Moreover, various hidden patterns and inconspicuous obscure trends in the data can be extracted using algorithms based on machine learning, multivariate statistics and  deep neural networks. We will enhance them with preconditioning algorithms to improve their effectiveness, enhance the robustness to the noise and artifacts.  We will characterize the primary relationships and patterns in the data, and extract secondary properties concealed in the data for use in AP and advisory decision support systems.Illinois Institute of Technology is a recipient of JDRF's research grant — Identification of Areas of Artificial Pancreas Algorithm Enhancements Through Big-Data Analysis — supported by Tidepool Big Data Donation Project.