Timely and consistent identification of crop diseases remains a persistent challenge in agricultural operations due to environmental variability, resource constraints, and the time-sensitive nature of intervention decisions. Although convolutional neural network (CNN)–based image classification systems have demonstrated high diagnostic accuracy in plant disease detection, their contribution to decision quality within decision support systems (DSS) has received comparatively limited analytical attention. This study adopts a secondary analytical research design to evaluate an AI-enabled rice disease identification system using validated performance data reported in prior empirical studies. No new experiments, model training, or data collection were conducted. Instead, reported classification accuracy and category-level performance are re-examined through a DSS lens, drawing on established theories of decision quality and information quality. Results from the source studies indicate high validation accuracy and robust performance under defined operating conditions. Interpreted from a decision support perspective, these findings suggest that reliable and consistent diagnostic outputs can reduce uncertainty and support timely intervention decisions, thereby enhancing decision quality, provided that decision-maker competence and system use conditions are appropriately aligned.
Castro et al. (Thu,) studied this question.