Machine Learning for Visualization and Prediction of Spatiotemporal Spread of COVID-19 in India

The application of GIS for disaster management and emergency response has provided the ease of producing meaningful information products that can otherwise be time and resource-intensive. Modeling, simulation, and visualization of GIS data provides disaster management decision-makers with the ease of using embedded information in effective knowledge generation and decision support processes, based on modeling geospatial data. Data science has emerged as a very promising interdisciplinary science that integrates various components that are related to data through a cycle that starts with problem definition, data acquisition, data processing, data exploration, and model selection and ends with product deployment. This data science life cycle is essentially the GIS modeling life cycle, which makes the process of using machine learning in GIS modeling a step forward. Artificial intelligence techniques, such as machine learning where the user applies a specific algorithm, allow the machine to learn from examples and experience to perform such complicated tasks. This makes the process of data exploration and product delivery worthwhile. This paper will provide a general overview of how supervised learning, unsupervised learning, as well as reinforced learning can be effective for disaster and emergency management applications, highlighting the autoregression integrated moving average (ARIMA) model for the prediction.