A Time Series Model for Short-term Prediction of Onchocerciasis among Farmers in Benue State-Nigeria
Ortaver Godwin Tyavyar
Department of Mathematics and Computer Science, Rev. Fr. Moses Orshio Adasu University, Makurdi, Benue State, Nigeria.
David Adugh Kuhe *
Department of Statistics, Joseph Sarwuan Tarka University, Makurdi, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Aims: The aim of this study is to apply time series modeling to predict short-term trends in Onchocerciasis infection among farmers in Benue State, Nigeria, and to generate reliable forecasts that can inform public health planning and agricultural productivity interventions.
Place and Duration of Study: The study was conducted in Benue State, Nigeria, using monthly Onchocerciasis infection data obtained from the Benue State Epidemiological Unit, Makurdi, covering the period from January 2009 to June 2025.
Methodology: The study employed descriptive statistics, the Ng-Perron modified unit root test, the Ljung-Box Q-statistic test, and the Autoregressive Moving Average (ARMA(p,q)) modeling framework. Model selection was based on information criteria and log-likelihood values, while forecast accuracy was evaluated using Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Theil’s Inequality Coefficient (TIC).
Results: Descriptive analysis revealed a mean monthly infection of 928 persons, with evidence of non-normality due to skewness and leptokurtosis. The Ng-Perron test confirmed stationarity of the series at level. The ARMA(3,3) model was identified as the best-fitting specification, explaining about 78.35% of the variation and exhibiting dynamic stability. Forecast evaluation showed strong predictive performance with low RMSE (0.2200), MAE (0.1883), MAPE (2.77%), and a near-zero TIC (0.0162). Analysis of model roots indicated an angular displacement of 44.75°, corresponding to an estimated Onchocerciasis life cycle of approximately eight months. Forecasts for July 2025 to June 2027 project an average monthly infection of about 865 persons (95% CI: 549-1,378), with a total of approximately 20,771 cases over the two-year period. The forecasts suggest that Onchocerciasis will persist at a steady level, underscoring the need for sustained drug distribution, strengthened vector control, continuous surveillance, and integrated health-agricultural policies to support effective disease control, protect farmers’ productivity, and improve food security in Benue State.
Conclusion: The study concluded that Onchocerciasis infection among farmers in Benue State exhibits a chronic and persistent pattern, with forecasts indicating a sustained public health burden. The ARMA(3,3) model provided a robust and reliable framework for monitoring and forecasting infection trends. Strengthening community-based surveillance, sustaining vector control and mass drug administration, integrating agricultural support programmes, enhancing farmer health education, and promoting intersectoral collaboration are essential for reducing the disease burden and its impact on postharvest crop losses in the state.
Keywords: ARMA model, onchocerciasis, time series analysis, stationarity, forecasting, stochastic processes, Nigeria