"background": "Persistent inefficiencies in water treatment yield within the country's infrastructure hinder reliable water supply. Existing operational models often lack robust predictive capabilities for long-term performance improvement, necessitating advanced analytical frameworks. ", "purpose and objectives": "This working paper presents a methodological evaluation of a novel time-series forecasting model designed to measure and predict yield improvement in water treatment systems. The objective is to assess the model's technical validity and applicability for infrastructure planning. ", "methodology": "The methodology employs an Autoregressive Integrated Moving Average with exogenous variables (ARIMAX) model, specified as \ Yt = \ + =1^{p\ \ Yt-i + =1^q\ -j + =1^r\ Xk, t + \, where Yt is treatment yield. Model parameters were estimated using maximum likelihood, with inference based on heteroskedasticity-robust standard errors. Evaluation utilised historical operational data from multiple facilities. ", "findings": "The model demonstrates a statistically significant positive trend in potential yield improvement, with a forecasted average increase of 12–18% over the projection period under optimal conditions. Parameter estimates for key maintenance and chemical dosing variables were significant at the 95% confidence level, indicating their strong predictive utility. ", "conclusion": "The ARIMAX framework provides a technically sound and operationally relevant method for forecasting yield improvements, offering a superior alternative to descriptive benchmarks for strategic asset management. ", "recommendations": "Implement the model as a decision-support tool for capital investment prioritisation and operational budgeting. Future work should integrate real-time sensor data to transition from periodic to continuous forecasting. ", "key words": "water treatment yield, time-series forecasting, ARIMAX, infrastructure performance, asset management", "contribution statement": "This paper provides a novel, rigorously evaluated forecasting methodology that enables quantitative, evidence
Ochieng et al. (Thu,) studied this question.