Scholarlink

Background <p>Ethiopia has been faced with the continual resurgence of malaria. It affects the health of the young workforce, which is believed to affect and slow economic growth.</p> Objective <p>Analyze and forecast the incidence of malaria in the next years (2026–2030) on the basis of historical data from Bahir Dar city in the Amhara National Regional State of Ethiopia.</p> Methods <p>A forecasting framework—seasonal autoregressive integrated moving average with exogenous factors (SARIMAX) model—was developed using malaria data from Amhara regional health bureau and exogenous regressors, weather data, from World Weather Online. The dataset comprising 90 monthly data points, spanning from January 1, 2018 to June 30, 2025, was split to develop and validate the model, reserving the first 80% (January 1, 2018–December 31, 2023) for training the model and the final 20% (January 1, 2024–June 30, 2025) for testing forecasting performance. We used the fitted model to forecast for the next 5&#xa0;years using Python version 3.11.</p> Results <p>The SARIMAX (1, 2, 2) (1, 2, 2, 12, exog) model, with weather data as exogenous regressors, fit the historical data well. It revealed an increasing trend, as evidenced by the in-sample fit, out-of-sample forecast and future prediction values, which consistently increased over the prediction horizon. None of the weather condition data showed a statistically significant predictive relationship with malaria incidence (<i>p</i> &gt; 0.05). The evaluation metrics, mean absolute percentage error (MAPE), confirmed reasonable predictive accuracy (28.3%).</p> Conclusion <p>Our study demonstrates an upward trend in forecasted malaria cases for the upcoming years, suggesting a potential breakdown in current strategies. The result underscores the necessity of a targeted, localized early warning system to manage resource allocation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Forecasting malaria incidence in a resource-limited urban setting with climate variables as exogenous regressors: time series analysis using a SARIMAX model in Bahir Dar, Ethiopia

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月12日

查看全文
  • Tesfaye Taye Gelaw,
  • Meseret Addisu Abera

摘要

Background

Ethiopia has been faced with the continual resurgence of malaria. It affects the health of the young workforce, which is believed to affect and slow economic growth.

Objective

Analyze and forecast the incidence of malaria in the next years (2026–2030) on the basis of historical data from Bahir Dar city in the Amhara National Regional State of Ethiopia.

Methods

A forecasting framework—seasonal autoregressive integrated moving average with exogenous factors (SARIMAX) model—was developed using malaria data from Amhara regional health bureau and exogenous regressors, weather data, from World Weather Online. The dataset comprising 90 monthly data points, spanning from January 1, 2018 to June 30, 2025, was split to develop and validate the model, reserving the first 80% (January 1, 2018–December 31, 2023) for training the model and the final 20% (January 1, 2024–June 30, 2025) for testing forecasting performance. We used the fitted model to forecast for the next 5 years using Python version 3.11.

Results

The SARIMAX (1, 2, 2) (1, 2, 2, 12, exog) model, with weather data as exogenous regressors, fit the historical data well. It revealed an increasing trend, as evidenced by the in-sample fit, out-of-sample forecast and future prediction values, which consistently increased over the prediction horizon. None of the weather condition data showed a statistically significant predictive relationship with malaria incidence (p > 0.05). The evaluation metrics, mean absolute percentage error (MAPE), confirmed reasonable predictive accuracy (28.3%).

Conclusion

Our study demonstrates an upward trend in forecasted malaria cases for the upcoming years, suggesting a potential breakdown in current strategies. The result underscores the necessity of a targeted, localized early warning system to manage resource allocation.