Convolution neural network based infection transmission analysis on Covid-19 using GIS and Covid data materials

Area of Science:

• Epidemiology
• Data Science
• Environmental Science

Background:

• Accurate prediction of infectious disease transmission is crucial for public health interventions.
• Understanding the influence of environmental and mobility factors on disease spread is complex.

Purpose of the Study:

• To introduce a novel Convolutional Neural Network-based model, CNN-CITA, for analyzing and predicting COVID-19 transmission.
• To enhance the identification of vulnerable sources for targeted intervention strategies.

Main Methods:

• Integration of climate data (temperature, humidity, rainfall) from remote sensing with Global Information System (GIS) data on population locations.
• Utilizing deep learning networks to analyze merged datasets, focusing on person location and mobility patterns across different timeframes.
• Calculating the Infection Transmission Rate (ITR) for distinct regions and time windows (before and after movement).

Main Results:

• The CNN-CITA model effectively predicts vulnerable sources based on infection rates and ITR values.
• The model demonstrates high performance in identifying areas at higher risk of infection transmission.
• Identified vulnerable sources aid in supporting the reduction of overall infection rates.

Conclusions:

• CNN-CITA offers an advanced approach to COVID-19 transmission analysis by integrating diverse datasets.
• The model's ability to predict vulnerable sources can significantly aid public health efforts in controlling disease spread.
• This methodology provides a robust framework for improving infectious disease transmission modeling.