FORECASTING OF COVID-19 PANDEMIC IN NIGERIA USING REAL STATISTICAL DATA

In this paper, we used data released by Nigeria Center for Disease Control (NCDC) every 24 hours for the past consecutive two months to forecast the Coronavirus disease 2019 (COVID-19) cases for the months (September – October 2020). The linear regression forecasting model and R software package are used for the forecast and simulations respectively. The COVID-19 cases in Nigeria is on a decreasing trend and the forecast result show that in the next two months, there is going to be a decrease in new COVID-19 cases in Nigeria. COVID-19 in Nigeria can be drastically reduced if the organizations, management, government or policymakers are constantly proactive 2 ABIOYE, UMOH, PETER, EDOGBANYA, OGUNTOLU, KAYODE, AMADIEGWU concerning these research findings.


INTRODUCTION
The coronavirus disease 2019 (COVID-19) is a highly contagious respiratory disease triggered by a strain of coronavirus that causes illness in human. The disease spreads from person to person via infected air droplets that are projected during sneezing or coughing. It may also be transmitted when humans have contact with hands or objects containing the virus, and when infected hands touch their eyes, nose or mouth. In December 2019, COVID-19 was first identified in China, but has now spread around the world.
COVID-19 was first announced in Nigeria on February 27, 2020 [1][2][3], and since then, it has been a major concerned to public health organizations. Presidential Task Force (PTF) was constituted as a matter of urgency to liaise with Federal Ministry of Health (FMOH) and Nigeria Center for Disease Control (NCDC) to give situation reports every 24 hours of COVID-19 new cases in Nigeria. Therefore, we consider the data given daily from the situation reports by NCDC on COVID-19 new cases in Nigeria from June 1, 2020 to August 31, 2020. These data of COVID-19 cases in Nigeria can be used by researchers to determine how the disease can be control. The data can also be used to know how fast the spread of COVID-19 in Nigeria and the strategies measure that need to be taken so as curb the spread of the disease.

EXPERIMENTAL DESIGN, MATERIALS AND METHODS
The data used for this research were obtained from the Nigeria Centre for Dieases Control (NCDC) the data covers a period of three (3) months, starting from June 1, 2020 to the August 31, 2020.
The data were collected daily from the NCDC website for a period of 92 days. The first day of COVID-19 occurrence in Nigeria was on 27 th of February 2020, and up to the date of this report, the NCDC have provided an up to date data on the COVID-19 pandemic. The first death case

2.1
Arima model An autoregressive model of order p can be written as where t  is white noise. This is like a multiple regression but with lagged values of t y as predictors. We refer to this as an AR(p) model, an autoregressive model of order p [4]. Rather than using past values of the forecast variable in a regression, a moving average model uses past forecast errors in a regression-like model.
where c is white noise. We refer to this as an MA(q) model, a moving average model of order q.
Of course, we do not observe the values of t  , so it is not really a regression in the usual sense.

The prophet package
The Prophet package uses a decomposable time series model with three main model components: trend, seasonality and holidays. They are combined in the following equation is a linear or logistic growth curve, s(t) represents periodic changes, h(t) takes care of holiday effects and t  is an error term which accounts for any unusual changes not accommodated by the model. Therefore, using the above models, the following graphs and tables are obtained.          In recent months, the COVID-19 pandemic in Nigeria has been on a decreasing trend and this is shown in figure 1. The COVID-19 dataset is shown graphically in figure 1. For effective forecasting of the COVID-19 data, the data was stored as a CSV file which was then loaded into ARIMA model and the Prophet package. Figure 2 shows the general trend and weekly trend of the COVID-19 dataset for the period of study. These cases are then compare with the official cases that were published by NCDC within that period. Figure 6 shows the autocorrelation function graph while figure 7 shows the partial autocorrelation graph of the COVID-19 dataset.

DISCUSSIONS OF RESULTS
These plots gave insight on the best ARIMA (0, 1, 1) model to use for the forecast. September 2020 together with its lower and upper boundaries is shown in Figure 5. This graph is plotted as an extension of the raw data. Figure 3 shows upper and lower forecast values for September 2020 compared with the New cases published by NCDC within the first 21 days of September 2020. Table 3 shows predicted values of COVID-19 for the month of September and October 2020.

CONCLUSION
The forecast of Covid 19 cases in Nigeria has been carried out and the following conclusions are presented. The plotted data used shows that the Covid 19 cases in Nigeria is on the decline. This research also shows that the trend in the Covid cases within the study period (1 st