Hiring in life sciences? Share your open positions with our professional community. Read more Close

Advertisement

Application of Data Science for COVID-19 Modeling in Africa: A Comparative Study of ARIMA and ANN Models, a Retrospective Time-Series Analysis.

Created on 12 Nov 2026

Authors

Kofi Afriyie Nyamekye, Eric Sampane-Donkor, Josephine Aboagye, Felicity Bennin, Haruna Latif, Justice Moses K Aheto

Published in

Health science reports. Volume 9. Issue 11. Pages e73498. Epub Nov 10, 2026.

Abstract

The COVID-19 pandemic highlighted the urgent need for robust forecasting tools to guide public health responses in Africa, where data infrastructure remains limited, and epidemic dynamics differ substantially from high-income settings.
Daily COVID-19 surveillance data were obtained from the WHO COVID-19 Dashboard for Egypt, Ghana, Kenya, Morocco, Nigeria, South Africa, and Tunisia. ARIMA models were fitted using the auto.arima function in R, with stationarity assessed using Augmented Dickey-Fuller (ADF) and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) tests. ANN models were implemented using the nnetar algorithm, configured with lag parameter p = 7, seasonal lag p = 1, hidden layer size of 10 nodes, and 20 repeats. Model accuracy was evaluated using root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and R-squared (R 2).
ARIMA outperformed ANN for four countries (Egypt, Kenya, Morocco, Nigeria), while ANN performed better for three countries (Ghana, Tunisia, South Africa). Egypt achieved the highest forecast accuracy (MAPE = 0.09% for ARIMA). South Africa presented the greatest forecasting challenge due to high volatility and multiple distinct waves, though ANN substantially outperformed ARIMA in this setting, reducing RMSE by 56% and MAPE by 51%. Negative R 2 values for six of seven countries reflected the structural inability of backward-looking autoregressive models to anticipate variant-driven wave peaks, consistent with limitations documented in COVID-19 forecasting literature.
Neither model is universally superior; model selection should be context-dependent. ANN is recommended for settings with stable case fatality rates and complex non-linear dynamics (e.g., Ghana) and for high-volume settings with multiple distinct waves (e.g., South Africa). ARIMA is recommended for moderate-transmission settings with more linear patterns where interpretability is prioritized.

PMID:
43016922
Bibliographic data and abstract were imported from PubMed on 12 Nov 2026.

Read full publication at:
Please sign in to see all details.

Advertisement

Stats

  • Community rating n/a 0 votes
  • Reviewers' rating n/a 0 votes
  • Your rating

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

  • Recommendations n/a n/a positive of 0 vote(s)
  • Views 7
  • Comments 0

Recommended by

  • No recommendations yet.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.

Advertisement