Traceable Organization of Food Quality‐ and Safety‐Related Postharvest Evidence in Fresh Tomato Literature

ABSTRACT

Postharvest evidence relevant to food quality and safety in fresh tomato literature is often scattered across result sections, tables, and figure captions, making condition‐specific comparison and evidence trace‐back difficult. This study developed and evaluated a traceable approach for organizing such evidence from the literature. The approach combined evidence segmentation, LLM‐assisted observation recovery, rule‐guided normalization, human review, and structured organization through a Source Article–Evidence Unit–Observation framework. Evaluation was conducted on a small controlled corpus of five English full‐text articles, which yielded 17 evidence units and 24 reference observations. The results showed stable performance in identifying evidence units that contained valid observations, whereas field extraction was less consistent and normalization was comparatively robust. The organized observation records supported evidence trace‐back and condition‐specific comparison across storage‐, transport‐, packaging‐, maturity‐, and treatment‐related contexts, including observations on firmness, color, weight loss, lycopene, sugar content, microbial count, and pH. Within this controlled setting, the findings support the feasibility of organizing food quality‐ and safety‐related postharvest evidence in fresh tomato literature into traceable and reviewable observation records for literature‐based comparison and interpretation. These findings should be understood within the limits of the present small‐scale validation and should not yet be generalized to a broader literature domain.