MACHINE LEARNING BASED MODEL FOR THE PREDICTION OF COVID-19 PANDEMIC SPREAD LEVEL IN ETHIOPIA

Abstract

The Novel Corona Virus Disease (COVID-19) is an epidemic that first broke out in Wuhan, China, in December 2019 and has spread worldwide, identified as leading to an ongoing pandemic. The spread of the outbreak increases a major national as well as an international crisis and learning influences the most important aspect of life and disturbs the political, social, economic, religious, and financial structure of the globe. Ethiopia is one country that holds 1.47% of the world population size and also it is a country that is affected by coronavirus pandemic. Over recent years, machine learning has turned very reliable in the medical field for the prediction and diagnosis in the health sector. The proposed study aimed to automate the prediction of covid-19 pandemic spread level in Ethiopia using a machine learning-based model that helps to predict the long-term (30, 60, 100, and 150 days) easily and inform the country will be able to tackle this pandemic on time. And for this work, the researchers had collected secondary time series data from Ethiopian Health Institute National Data Management Center, time-series data from March 13, 2020, to Feb 3, 2022, to analyze and design the machine learning model by using python programming with Spyder development environment. And study focus on proposed models like Long Short Term Memory(LSTM), Multilayer Perceptron(MLP), Support Vector Regression(SVR), Polynomial Regression(PR), and Random Forest Regression(RFR) and the models evaluated using regression models evaluation metrics such as mean squared error(MSE), mean absolute error(MAE), R2 scored(R2 ), and root mean squared error(RMSE). According to above evaluation metrics; LSTM model performed well with less error. The LSTM model evalution result in conformed case( MAE=832.750, R2 =0.999); recovered case(MAE=562.542, R2 =0.999); and in death case prediction(MAE=10.02137, R 2 =0.999) this shown that the better performance of the model among proposed models in the study. The result produce by the study promising that LSTM good to predict current scenarios of covid-19 in Ethiopia. The prediction of model has the capability to predict for the next five months.