Explainable Artificial Intelligence Methods for Analysis of Factors Influencing Covid-19 Cases in Türki̇ye

Abstract COVID-19 disease emerged in Wuhan, China, in 2019 and evolved into a pandemic that negatively affected all countries worldwide. Researchers have employed methods such as machine learning and spatial machine learning methods, including spatial and multiple linear regression, geographically weighted regression and geographical random forests, to determine the importance of factors such as sociocultural, demographic, environmental, racial, and economic development related to COVID-19 cases or deaths caused by COVID-19. In this study, the explainability of the general factors analyzed by the random forest model of COVID-19 cases based on provinces in Türkiye was examined. In addition, traditional machine learning methods, compared with spatial machine learning models, as the Explainable Artificial Intelligence (XAI) methods directed toward employing spatial associations. The most important factors in the model might differ locally among provinces according to cluster. The spatial machine learning models performed better than the random forest model.