With Machine Learning Methods Covid-19 Investigation of Data: The Case of Turkey

An Analysis of Covid-19 Data With Machine Learning Methods: The Case of Turkey

Abstract

Covid 19 virus is one of the most important problems affecting our health and life today. It is considered that the effect of this virus for a normal patient continues approximately one month. Turkey's Health Ministry declared that the daily cases of death, recovering patient, tests and the number of seriously ill and these cases aimed to evuluate on a monthly basis the progress of virüs. In our research were used to announced between March 2020 and March 2021 the data set twelf montly. This data set was analysed with the Random Forest algorithmn, which is one of the machine learning classification methods. As a result of the analysis, the method was tested with Precision, Recall, Score F and AUC performance criteria. In addition, the importance of the variables used for the model was evaluated.  As a result of the analysis, the accuracy (OOB) of our model was found to be 83%. Performance criteria were found to have an precision rate of 90%, a recall rate of 89%, an F score of 89%, and an area under the curve (AUC) of 99%.   The most important variable for the accuracy of the variable significance model was the Daily Healing Number, while the most important variable in determining the grade was the Daily Healing Patient Number.