@article{HIJAB HASSAN_2023, title={Assessing The COVID-19 Trends In Pakistan Using Predictive Machine Learning Techniques: An Empirical Study}, volume={3}, url={https://journals.pakistanreview.com/index.php/GJPR/article/view/173}, abstractNote={<p><span dir="ltr" style="left: 18.12%; top: 20.87%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.990862);" role="presentation">The Coronavirus (also referred to as COVID-19) which started in Wuhan, China </span><span dir="ltr" style="left: 18.12%; top: 23%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.954008);" role="presentation">on December 2019 has taken the world by storm. Scientists across the globe have </span><span dir="ltr" style="left: 18.12%; top: 25.12%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.02761);" role="presentation">used epidemiological models to predict the spread of the virus along with the </span><span dir="ltr" style="left: 18.12%; top: 27.25%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.97526);" role="presentation">death rate and make different outbreak predictions. Also, prediction models have </span><span dir="ltr" style="left: 18.12%; top: 29.38%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.00795);" role="presentation">been utilized for new policies to control the spread of the virus.</span> <span dir="ltr" style="left: 75.21%; top: 29.38%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.949432);" role="presentation">Because of the </span><span dir="ltr" style="left: 18.12%; top: 31.5%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.985596);" role="presentation">complex and irregular nature of the virus, it has been hard to forecast the trends </span><span dir="ltr" style="left: 18.12%; top: 33.63%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.967726);" role="presentation">in different nations specially using conventional mathematical models such as the </span><span dir="ltr" style="left: 18.12%; top: 35.76%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.964383);" role="presentation">SIR (Susceptible Infected resistant) model. Therefore, this study analyzes the five </span><span dir="ltr" style="left: 18.12%; top: 37.89%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.999728);" role="presentation">waves of COVID-19 that have hit Pakistan since February 2020 using Machine </span><span dir="ltr" style="left: 18.12%; top: 40.01%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.965783);" role="presentation">Learning models. Advanced predictive models Predictive models such as Logistic </span><span dir="ltr" style="left: 18.12%; top: 42.14%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.99716);" role="presentation">Growth and Autoregressive Integrated Moving Average model (ARIMA) are uti</span><span dir="ltr" style="left: 18.12%; top: 44.27%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.980849);" role="presentation">lized for predicting and modeling contagion spread trends. The study uses these </span><span dir="ltr" style="left: 18.12%; top: 46.4%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.968741);" role="presentation">models to capture the variation in the incidence of daily cases in each province of </span><span dir="ltr" style="left: 18.12%; top: 48.52%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.993507);" role="presentation">Pakistan. The time-series data utilized for this study is obtained from the official </span><span dir="ltr" style="left: 18.12%; top: 50.65%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.949641);" role="presentation">website of the government of Pakistan; consisting of daily caseload for each region </span><span dir="ltr" style="left: 18.12%; top: 52.78%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.987228);" role="presentation">of Pakistan. There are two main contributions of the paper: first, it compares the </span><span dir="ltr" style="left: 18.12%; top: 54.9%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.967321);" role="presentation">modeling accuracy of two widely used disease growth models ARIMA, and Logis</span><span dir="ltr" style="left: 18.12%; top: 57.03%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.01789);" role="presentation">tic Growth Model, in the case of Pakistan.</span> <span dir="ltr" style="left: 56.96%; top: 57.03%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.975287);" role="presentation">Secondly, it recommends the model </span><span dir="ltr" style="left: 18.12%; top: 59.16%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.971859);" role="presentation">better suited for datasets similar to Pakistan, which have fluctuations in numbers. </span><span dir="ltr" style="left: 18.12%; top: 61.29%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.00139);" role="presentation">One of the main limitation of this research is that although, one solution for this </span><span dir="ltr" style="left: 18.12%; top: 63.41%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.985491);" role="presentation">uncertainty has been the use of Machine Learning predictive techniques, limited </span><span dir="ltr" style="left: 18.12%; top: 65.54%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.00977);" role="presentation">data is available for Pakistan.</span> <span dir="ltr" style="left: 45.96%; top: 65.54%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.02226);" role="presentation">The findings of this research indicated that the </span><span dir="ltr" style="left: 18.12%; top: 67.67%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.984581);" role="presentation">logistic model could not model everyday COVID-19 case numbers effectively for </span><span dir="ltr" style="left: 18.12%; top: 69.8%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.982419);" role="presentation">the overall pandemic wave, and the model tried to decrease the error, producing </span><span dir="ltr" style="left: 18.12%; top: 71.92%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(1.00725);" role="presentation">an inaccurate plot.</span> <span dir="ltr" style="left: 36.05%; top: 71.92%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.993765);" role="presentation">However, it showed better results when the waves were di</span><span dir="ltr" style="left: 18.12%; top: 74.05%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.968927);" role="presentation">vided into smaller sections. The RMSE were less compared to the ARIMA Model. </span><span dir="ltr" style="left: 18.12%; top: 76.18%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.979959);" role="presentation">Lastly, the researcher also recommends other models which could be utilized for </span><span dir="ltr" style="left: 18.12%; top: 78.3%; font-size: calc(var(--scale-factor)*11.96px); font-family: sans-serif; transform: scaleX(0.996632);" role="presentation">further modeling of COVID-19 trends in Pakistan.</span></p>}, number={1}, journal={Graduate Journal of Pakistan Review (GJPR)}, author={HIJAB HASSAN}, year={2023}, month={Mar.} }