Pandemic Outbreak Prediction using Optimization-based Machine Learning Model

Abstract

Many pandemic epidemics have a variety of effects on people. Disease modelling is important in order to predict and evaluate the effects of these pandemics. A variety of statistical and machine learning (ML) models are built to produce the forecast. The models employ several ML strategies, but due to the available dataset, they are unable to achieve higher accuracy. To avoid the problem and improve forecast accuracy, we suggested an ML-based prediction model. This study optimises the parameters of the already employed machine learning models using the proposed Ant Colony Optimization approach (ACO). A comparison of different machine learning (ML) approaches, such as Polynomial Regression (PR), Support Vector Machine (SVM), and Linear Regression, is provided to predict the pandemic outbreak (LR). For COVID-19 datasets, the accuracy and Root Mean Square Error (RMSE) score of the proposed model are used to assess its performance. The results show that, as assessed by the RMSE score, the suggested method delivers good accuracy for daily prediction. The outcome forecast shows that PR-ACO outperforms other ML strategies in terms of final results. According to the results of the predictions, the proposed ACO parameter optimising algorithm will increase the capacity of current ML techniques to anticipate outbreaks across diverse countries.