Abstract:
Severe Acute Respiratory Syndrome is the primary cause of the current pandemic coronavirus disease (COVID-19).
The first case was reported in Wuhan, China, on December 30th, 2019 with the first case on 13thMarch, 2020 in Kenya. This
contagious disease has become a global issue because it has resulted in millions of deaths, economic disruption leading to loss
of employment and economic instability. Researchers have fitted time series models but using a short data length and without a
transition. There was therefore a need to model a longer data period of daily COVID-19 cases with a transition in Kenya using
theAutoregressiveIntegrated Moving Average (ARIMA) model and forecast. Secondary data from the World Health
Organization from 13thMarch, 2020 to 30thApril, 2023 was analyzed using R software. The data was found to be non-stationary
using the Augmented Dickey Fuller test and regular differencing was done to make it stationary. The Box-Jenkins methodology
was used to fit the model of the data and afterwards forecasting was done. The ARIMA (3,1,2) was selected as the best model
since it had the least Akaike Information Criterion and Bayesian Information Criterion among the possible models. Model
validation using test data was done by comparing the MAE, and RMSE of the model’s forecasts and it was the best amongst the
possible models with MAE = 2.77 and RMSE =2.88. The model was fitted to the daily COVID-19 data and forecasting was
then done for ninety days into the future
