The machine learning approach for predicting the number of intensive care, intubated patients and death: The COVID-19 pandemic in Turkey

Abstract

The coronavirus infection outbreak started in W u h a n city, China, in December 2019 and affected more than 200 countries in a year. Th e nu mber of p a ti ents dy in g from and infected with COVID-19 is increasing at an alarming rate in almost all affected countries. One of the most important factors in the COVID-19 death and case rates is the care of intensive care patients. The management of COVID-19 patients who need acute and/ or critical respiratory care has created a significant difficulty for h ealthcare systems worldwide. To prevent the further spread of COVID-19 around the world and to fight the disease, nonclinical computer-aided quick solutions such as artificial intelligence and machine learning are needed. Prediction techniques evaluate past situations and enable predictions about the future situation. In this study, using the dataset created from the data received from the World Health Organization and national database, the numbers of intensive care, intubated patients, and deaths from COVID-19 in Turkey were predicted by the random forest, bagging, support vector regression, classification and regression trees, and k-nearest neighbors machine learning regression methods. In this study, the random forest method has been the most successful algorithm for predicting the number of intensive care patients (r = 0.8698, RMSE = 188.5, MAE = 135.1, MAPE = 13%), the number of intubated patients (r = 0.9846, RMSE = 47.1, MAE = 39.7, MAPE = 9.2%), and the number of deaths (r = 0.9994, RMSE = 1.2, MAE = 0.9, MAPE = 3.5%). The results in this study, it has been shown that machine learning methods, which have been successfully applied in other epidemic diseases, will be successfully applied in the COVID-19 pandemic.

A study on next-generation digital tool for health data management: the e-Pulse portal

“…The last five years, the relationship between healthcare management and digital technology (El-Sherif et al ., 2022), blockchain (Aruna, 2021; Qu, 2022; Pekdemir, 2021; Katuu, 2022) and learning machine systems (Maldonado, 2021; Cihan, 2022) were increased. Accordingly, it can be said that digital transformation will accurate among healthcare management in the future.…”

A new approach for determining SARS-CoV-2 epitopes using machine learning-based in silico methods

“…Although epitope data are not yet available for SARS-CoV-2, since the gene and protein sequence information is known, the characteristics of the virus and the epitopes in the pathogen can be predicted by machine learning-based in silico methods ( Tahir ul Qamar et al, 2019 ). There are many studies in the literature on predicting the death of patients diagnosed with SARS-CoV-2 using machine learning/artificial intelligence algorithms, modeling the spread of infection ( Ceylan, 2020 , Cihan, 2022 ), diagnosis with medical image analysis ( Saygılı, 2021 ), and forecasting the COVID-19 vaccination rate ( Cihan, 2021 , Zhou and Li, 2022 ).…”

Classification of the death ratio of COVID-19 Pandemic using Machine Learning Techniques

“…There are many papers that implemented statistical methods to the COVID-19 dataset to build predictive models that can assess mortality rates [6,7,8]. Cihan (2022) estimated the number of intensive care, intubated patients and deaths caused by COVID-19 in Turkey with random forest, bagging, support vector regression, classification and regression trees and machine learning regression methods. As a result of the study, it was determined that the random forest algorithm produced the most successful results.…”

“…As a result of the study, it was determined that the random forest algorithm produced the most successful results. [9]. The models based on machine learning have been used for the prediction and classification of epidemic development [10].…”