¹Shanghai High School International Division, Shanghai, China 200231.
²California State Polytechnic University, Pomona, CA 91768.
³Amazon.com, Irvine, CA 92618.

---

Abstract—Gestational diabetes patients have to closely monitor their blood sugar levels four times a day using the traditional finger pricks, which often causes extra pains and inconvenience during the pregnancy. The monitoring approach without using finger pricks has not been widely used due to the low accuracy and high cost. In this project, we address this problem by using mobile computing and machine learning. A mobile app has been developed to collect the patient’s diet and the tested blood sugar level. Once a sufficient amount of data has been collected, the system is able to train the machine learning model and predict the patient’s blood sugar level based on the diet. Experiments show that our prediction without finger prick monitoring can reach to 91% accuracy when the patient is under a regular and routine diet with adequate daily exercises.

Index Terms—Gestational diabetes, machine learning, mobile computing.

[PDF]

Cite: Shiyu Sara Huang, Chou in Luk, Liping Zhou, Yu Sun, "An Intelligent Mobile System to Predict Blood Sugar Level for Gestational Diabetes Patients Using Machine Learning," Journal of Computers vol. 13, no. 10, pp. 1227-1234, 2018.