Utilizing fog computing and explainable deep learning techniques for gestational diabetes prediction

Abstract Gestational diabetes mellitus (GDM) is one of the pregnancy complications that poses a significant risk on mothers and babies as well. GDM usually diagnosed at 22–26 gestation. However, early prediction desirable it may contribute to decrease risk. The continuous monitoring for mother’s vital signs helps in predicting any deterioration during pregnancy. originality this paper provide comprehensive framework women monitoring. proposed Data Replacement Prediction Framework consists three layers which are: (i) IoT Layer, (ii) Fog (iii) Cloud Layer. first layer used IOT sensors aggregate sings from pregnancies using invasive noninvasive sensors. Then transmitted fog nodes processed finally stored cloud layer. main contribution located producing module implement two influential tasks Finding Methodology (DFM), Explainable Algorithm (EPM) DNN. First, DFM replace unused data free cache space new incoming items. replacement very important case healthcare system are frequent must be replaced continuously. Second, EPM predict incidence occur second trimester To evaluate our model, we extract 16,354 medical information mart intensive care (MIMIC III) benchmark dataset. For each woman, signs, demographic laboratory tests was aggregated. results model superior state art (ACC = 0.957, AUC 0.942). Regarding explainability, utilized Shapley additive explanation local global developed models. Overall, medically intuitive, allow with cost effective solution .