Predicting the Incidence of COVID-19 Using Data Mining

Abstract

Background: The high prevalence of COVID-19 has made it a new pandemic. Predicting both its prevalence and incidence throughout the world is crucial to help health professionals make key decisions. In this study, we aim to predict the incidence of COVID-19 within a two-week period to better manage the disease. Methods: The COVID-19 datasets provided by Johns Hopkins University, contain information on COVID-19 cases in different geographic regions since January 22 and are updated daily. Data from 252 such regions were analyzed as of March 29, 2020, with 17,136 records and 4 variables, namely latitude, longitude, date, and records . In order to design the incidence pattern for each geographic region, the information was utilized on the region and its neighboring areas gathered two weeks prior to the designing. Then, a model was developed to predict the incidence rate for the coming two weeks via a Least-Square Boosting Classification algorithm. Results: The model was presented for three groups based on the incidence rate: less than 200, between 200 to 1000, and above 1000. The model evaluation error rates were 4.71%, 8.54%, and 6.13%, respectively. Also, comparing the forecast results with the actual values in the period in question showed that the proposed model predicted the number of globally confirmed cases of COVID-19 with a very high accuracy of 98.45%. Conclusion: Using data from different geographical regions within a country and discovering the pattern of prevalence in a region and its neighboring areas, our boosting-based model was able to accurately predict the incidence of COVID-19 within a two-week period.

Predicting Mortality of COVID-19 Patients based on Data Mining Techniques

“…The highest accuracy in the present study was reported at 99.14% using the KNN2 algorithm. The accuracy of the regression model used in the study conducted by Ahouz et al [18] was obtained at 83% and in the study conducted by Li et al [21], it was obtained at 0.8698 with the SVM (Linear kernel) model. In the study conducted by Ahouz et al [18], 17136 records and 4 variables (latitude, longitude, history, and background) were analyzed.…”

“…The accuracy of the regression model used in the study conducted by Ahouz et al [18] was obtained at 83% and in the study conducted by Li et al [21], it was obtained at 0.8698 with the SVM (Linear kernel) model. In the study conducted by Ahouz et al [18], 17136 records and 4 variables (latitude, longitude, history, and background) were analyzed. Li et al [21] also identified stress symptoms in people due to the COVID-19 outbreak using data mining.…”

“…In reviewing the above studies, we found that none of these studies focused on predicting mortality in patients with COVID-19. These studies have been conducted with different aims of identifying patients with COVID-19 infection [13], predicting death in patients with suspected sepsis [22], and identifying stress symptoms in people due to outbreak of COVID-19 [21], and patients with high-risk COVID-19 [14] and predicting the rate of CO-VID-19 in the next two weeks [18]. Regarding the importance of predicting patients' mortality, Deschepper et al [23] stated that early prediction of mortality in hospitals can improve the patient outcomes and enable health care providers to take adequate and timely action to save lives.…”

“…Also, as mentioned in the studies [13,18,21], some different predictors have been used in the development of the models. All of these studies used fewer factors to develop their models compared to the present study.…”

“…Therefore, the neural network used in the present study consisted of 1 input, 11 nodes in the hidden layer, and 1 output. Samples of clinical data mining studies were cited to determine the hidden layers [17,18]. These studies emphasized the ability of two-layer neural networks.…”