Utility of Big Data in Predicting Short-Term Blood Glucose Levels in Type 1 Diabetes Mellitus Through Machine Learning Techniques

Abstract

Machine learning techniques combined with wearable electronics can deliver accurate short-term blood glucose level prediction models. These models can learn personalized glucose–insulin dynamics based on the sensor data collected by monitoring several aspects of the physiological condition and daily activity of an individual. Until now, the prevalent approach for developing data-driven prediction models was to collect as much data as possible to help physicians and patients optimally adjust therapy. The objective of this work was to investigate the minimum data variety, volume, and velocity required to create accurate person-centric short-term prediction models. We developed a series of these models using different machine learning time series forecasting techniques suitable for execution within a wearable processor. We conducted an extensive passive patient monitoring study in real-world conditions to build an appropriate data set. The study involved a subset of type 1 diabetic subjects wearing a flash glucose monitoring system. We comparatively and quantitatively evaluated the performance of the developed data-driven prediction models and the corresponding machine learning techniques. Our results indicate that very accurate short-term prediction can be achieved by only monitoring interstitial glucose data over a very short time period and using a low sampling frequency. The models developed can predict glucose levels within a 15-min horizon with an average error as low as 15.43 mg/dL using only 24 historic values collected within a period of sex hours, and by increasing the sampling frequency to include 72 values, the average error is reduced to 10.15 mg/dL. Our prediction models are suitable for execution within a wearable device, requiring the minimum hardware requirements while at simultaneously achieving very high prediction accuracy.

Forecasting of Glucose Levels and Hypoglycemic Events: Head-to-Head Comparison of Linear and Nonlinear Data-Driven Algorithms Based on Continuous Glucose Monitoring Data Only

“…Clearly, this last finding does not exclude that other nonlinear ML or DL techniques could change the picture (an exhaustive exploration of nonlinear techniques is practically impossible, also considering the number of new contributions constantly proposed in these fields), but proves that linear methods are still highly valuable options that offer an excellent trade-off between complexity and performance. It is worth noting that both the numerical and statistical findings of this analysis seem to be in line with most of the literature studies [ 14 , 15 , 16 , 33 , 38 , 41 , 42 , 51 ]. Nonetheless, we report a clear contrast with the findings in some other contributions [ 43 , 55 ].…”

“…Three ML models, successfully used in a wide range of regression problems, were considered: support vector regression (SVR) [ 39 , 40 ], regression random forest (RegRF) [ 41 ], and feed forward neural network (fNN) [ 42 ]. In addition, we considered a DL model, namely, long short-term memory (LSTM) network, which has shown promising results in glucose prediction [ 43 , 44 ].…”

Machine Learning Techniques for Hypoglycemia Prediction: Trends and Challenges

“…Quan et al [ 50 ], Dong et al [ 62 ], and Mhaskar et al [ 63 ] used neural networks trained on BG data for hypoglycemia prediction. On the other hand, Rodriguez et al [ 57 ] used three different ML models trained on BG data from 25 patients. A KRR-based system was presented by Marcus et al [ 58 ] while Seo et al [ 46 ] proposed another model to predict hypoglycemia that used BG values.…”

A Comparison of Feature Selection and Forecasting Machine Learning Algorithms for Predicting Glycaemia in Type 1 Diabetes Mellitus

“…Numerous attempts have been made to develop a reliable prediction of glucose in DM1 patients. In this case, there have been approaches from a univariate point of view [16], using Autoregressive Integrated Moving Average (ARIMA), Random Forest (RF) and Support Vector Machines (SVM) with acceptable results. However, although univariate approximations can be interesting in computationally restricted environments, multivariate methods have demonstrated higher accuracy [17].…”