چكيده به لاتين
From its inception, the COVID-19 has dramatically affected the lives of people around the world. The virus has caused many problems due to the unknown and lack of knowledge about this virus and in many cases has caused human deaths. Experience, on the other hand, shows that having an underlying disease, especially diabetes, exacerbates the disease and increases the risk of death. Proper and timely measures for Crohn's patients, especially those with an underlying disease such as diabetes, can prevent many of these complications and reduce mortality from the disease. One of these measures is the correct and timely diagnosis of the need for an intensive care unit. Unfortunately, due to the lack of facilities in hospitals in this situation, only a limited number of patients can be admitted to the intensive care unit, which makes the accurate and timely identification of patients who need this unit much more important. Machine learning is a subset of artificial intelligence that allows systems to learn and progress automatically. Many fields use machine learning, and the field of healthcare is no exception. With the help of machine learning tools, valuable patterns and rules in the field of health can be achieved that help speed up the healing process. However, finding these rules empirically is sometimes very time-consuming. In this study by using machine learning tools, the data related to 659 patients with coronavirus of Imam Khomeini Hospital in Tehran, some of whom were diabetic and some of whom were non-diabetic, are examined. The main purpose of this study is to create a highly accurate predictive model to predict the need of patients with COVID-19 in the intensive care unit by considering diabetes as an underlying disease. Finding important features in predicting the need for an intensive care unit for these patients for both diabetic and non-diabetic groups is another goal of this study. Finally, the Gradient Boosting algorithm performed best for both groups, and with relatively good accuracy, we were able to provide a model for predicting the need for an intensive care unit for both groups. Finally, Neutrophil, Lymph, Abnormal lung signs, Urea, and calcium for diabetic patients with coronavirus, as well as Neutrophil, Lymph, WBC, Urea, and Sputum for non-diabetic patients with coronavirus were identified as important features to predict the need for an intensive care unit.
