Methodological Evaluation and Time-Series Forecasting for Manufacturing Yield Improvement in Ethiopia (2000–2026)

Manufacturing yield is a critical metric for industrial efficiency and economic development. In many developing economies, systematic evaluation of production systems and reliable forecasting of yield trends are lacking, hindering targeted improvement strategies. This study aims to methodologically evaluate manufacturing plant systems and develop a robust time-series forecasting model to predict yield improvement trajectories. The objective is to provide an evidence-based tool for strategic planning within the industrial sector. A hybrid methodology was employed, integrating system performance evaluation across multiple plants with statistical forecasting. The core forecasting utilised an Autoregressive Integrated Moving Average (ARIMA) model, specified as Yₜ = + ₁ Yₓ-₁ + ₁ ₓ-₁ + ₜ, where parameters were estimated via maximum likelihood. Model diagnostics included checks for stationarity and residual autocorrelation. The model forecasts a positive but decelerating annual yield improvement trend, with a projected increase of approximately 2. 7% per annum over the forecast horizon. Parameter estimates were statistically significant, and the model's mean absolute percentage error (MAPE) was 3. 2%, indicating robust predictive accuracy within a 95% confidence interval. The methodological framework successfully integrates system evaluation with quantitative forecasting, demonstrating that sustained yield gains are achievable but require continued technological and process interventions to maintain momentum. Industry stakeholders should adopt similar integrated evaluation and forecasting approaches for capacity planning. Investment should be prioritised in process automation and workforce skill development to address forecasted deceleration. manufacturing yield, time-series analysis, ARIMA modelling, industrial forecasting, system evaluation, production efficiency This paper provides a novel integrated framework combining systematic plant evaluation with a statistically robust forecasting model, yielding a practical tool for evidence-based industrial policy and management.