Evaluation of COVID-19 Diagnosis Systems Using Machine Learning

Authors

Mojtaba Azizi Faculty of Chemistry and Chemical Engineering, Malek Ashtar University of Technology, Author
Saeed Alikhani Department of Computer Engineering, North Tehran Branch, Islamic Azad University, Tehran, Iran Author

Keywords:

Data mining, Diagnosis, COVID-19, Artificial intelligence, Kaggle

Abstract

Data mining is the process of analyzing large volumes of data to uncover meaningful information. This process is employed by companies to transform raw data into valuable insights, as well as to explore and analyze vast amounts of information to identify significant patterns and trends. Data mining aids companies in solving problems, reducing risks, and capitalizing on new opportunities. This branch of data science derives its name from the analogy of searching for valuable information in a large database, similar to mining ore from a mountain.
The COVID-19 virus has infected over 100 million people and resulted in nearly three million deaths worldwide. To curb the spread of this unprecedented pandemic, increasing attention has been directed toward the use of artificial intelligence techniques to develop tools that assist physicians in various tasks. Despite promising results for diagnostic tasks (such as COVID-19 detection), the development of prognostic models—whether for predicting ICU admissions, other outcomes (including mortality), or classifying patients based on risk—has so far lagged behind in the context of COVID-19 diagnosis.

In the first chapter, the relevant data mining techniques are explored. In the second chapter, data mining operations are conducted on a COVID-19 dataset obtained from Kaggle, and the results are reported. Finally, in the third chapter, studies related to COVID-19 diagnosis are compared.