Forecasting COVID-19 new cases in Algeria using Autoregressive fractionally integrated moving average Models (ARFIMA)

In this research, an ARFIMA model is proposed to forecast new COVID-19 cases in Algeria two weeks ahead. In the present study, public health database from Algeria health ministry has been used to build an ARFIMA model and used to forecast COVID-19 new cases in Algeria until May 11, 2020.

Background

The aim of this study is first to find the best prediction method among the two techniques used and type of memory, either short or long, of the model constructed for the daily confirmed cases in Algeria, then make forecasts of the confirmed cases in the fifteen next days.

Methods

This study was conducted based on daily new cases of COVID-19 that were collected from the official website of Algerian Ministry of Health from March 1, 2020 to April 26, 2020. Auto Regressive Integrated Moving Average (ARFIMA) model was used to predict the trend of confirmed cases. The evaluation of the fractional differentiation parameter ( d ) is carried out using OxMetrics 6 software.

Results

The ARFIMA model (0, 0.431779, 0) build for Algeria, has a long memory and an upward trend over the next fifteen days and which coincides with the holy month of Ramadhan.

Conclusions

The forecasted results obtained by the proposed ARFIMA model can be used as a decision support tool to manage medical efforts and facilities against the COVID-19 pandemic crisis.