A Hybrid CNN and GRU-based Spatial-temporal Marburg Virus Disease Hotspot Association Mining for Health Management in Kenya

The paper demonstrates the analysis of Marburg Virus Disease (MVD) hotspot based on a New CNN_GRU based Association Rule Mining (ARM) model. In this work, a new technique is used for classification and prediction of MVD hotspot in Kenya. The main contribution of the research includes a combination of the CNN-GRU to identify the disease with minimum features. ARM is used to highlight the hotspots in Kenya. Maximum Entropy (MaxEnt) distribution is used to obtain uniform distribution among the training data from all the states to predict the hotspot. The performance of combined CNN-GRU model is compared with the existing algorithms. The New combined model gives an Accuracy of 99.1%, Precision of 98.8%, Recall of 98.9% and F1-Score of 98.4%.