Evaluating data sufficiency feasibility and robustness in resource aware forecasting for sustainable decisions

Abstract Sustainable decision systems often operate with limited historical data, measurement channels, computation, and imperfect records, yet forecasting methods are commonly compared mainly by predictive accuracy under full information. We develop and evaluate a resource-aware forecasting framework that treats predictive risk, finite-grid first-crossing data sufficiency, measurement burden, feasibility, uncertainty, missingness robustness, cross-wave change, and computational cost as distinct evaluation dimensions. The empirical study uses the Uganda component of the World Bank Living Standards Measurement Study–Integrated Surveys on Agriculture harmonised longitudinal panel. Household-wave transitions 1–2 through 4–5 are used for fitting, transition 5–7 for calibration and threshold setting, and transition 7–8 for held-out future-wave evaluation. Three scalar tasks predict next-wave household dietary diversity, log per-capita consumption, and log agricultural harvest value, while a fourth predicts an 11-component crop-composition vector constrained to the probability simplex. Predictive accuracy alone can conceal important deployment failures. For Ridge dietary-diversity forecasts, only four predictions violate the admissible range, yet they account for 69.1% of total raw squared error; valid projection reduces root mean squared error from 3.720 to 2.071. For crop composition, simplex projection reduces Ridge mean squared Euclidean loss from 0.345 to 0.294, with pointwise non-worsening in all 2227 future-wave cases up to numerical precision. Measurement reductions can be inexpensive for some tasks, while missingness effects depend strongly on mechanism. Chronologically calibrated 90% residual intervals attain empirical future-wave coverage of approximately 0.918–0.955. These findings support resource-aware forecasting as an evaluation discipline in which accuracy, feasibility, robustness, uncertainty, and operational burden are assessed jointly under a declared operating regime.

Authors

Institutions

Publication Details

Journal
Discover Sustainability
Published
2026-11-16
DOI
https://doi.org/10.1007/s43621-026-05248-5
Primary Topic
Forecasting Techniques and Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Evaluating data sufficiency feasibility and robustness in resource aware forecasting for sustainable decisions

Ronald Katende
Discover Sustainability
Forecasting Techniques and Applications
article

Evaluating data sufficiency feasibility and robustness in resource aware forecasting for sustainable decisions

Ronald Katende
article en

Abstract

Abstract Sustainable decision systems often operate with limited historical data, measurement channels, computation, and imperfect records, yet forecasting methods are commonly compared mainly by predictive accuracy under full information. We develop and evaluate a resource-aware forecasting framework that treats predictive risk, finite-grid first-crossing data sufficiency, measurement burden, feasibility, uncertainty, missingness robustness, cross-wave change, and computational cost as distinct evaluation dimensions. The empirical study uses the Uganda component of the World Bank Living Standards Measurement Study–Integrated Surveys on Agriculture harmonised longitudinal panel. Household-wave transitions 1–2 through 4–5 are used for fitting, transition 5–7 for calibration and threshold setting, and transition 7–8 for held-out future-wave evaluation. Three scalar tasks predict next-wave household dietary diversity, log per-capita consumption, and log agricultural harvest value, while a fourth predicts an 11-component crop-composition vector constrained to the probability simplex. Predictive accuracy alone can conceal important deployment failures. For Ridge dietary-diversity forecasts, only four predictions violate the admissible range, yet they account for 69.1% of total raw squared error; valid projection reduces root mean squared error from 3.720 to 2.071. For crop composition, simplex projection reduces Ridge mean squared Euclidean loss from 0.345 to 0.294, with pointwise non-worsening in all 2227 future-wave cases up to numerical precision. Measurement reductions can be inexpensive for some tasks, while missingness effects depend strongly on mechanism. Chronologically calibrated 90% residual intervals attain empirical future-wave coverage of approximately 0.918–0.955. These findings support resource-aware forecasting as an evaluation discipline in which accuracy, feasibility, robustness, uncertainty, and operational burden are assessed jointly under a declared operating regime.

Discover Sustainability
Kabale University (UG)
Zero hunger, Climate action
Forecasting Techniques and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon