Combining correlation analysis and machine learning to identify indicators strongly linked to quality variations in low-salt fermented shrimp sauce during storage.

Developing reduced-salt shrimp sauce has become an inevitable trend, while increases the risk of spoilage during its storage. The current quality control criteria for shrimp sauce fails to provide accurate and holistic quality assessment. The establishment of new quality control standards is need and necessitates clarifying the quality evolution patterns of low-salt shrimp sauce. During storage, pH, total volatile basic nitrogen (TVB-N), histamine, tartaric acid, succinic acid, hypoxanthine and aromatic compounds increased significantly, while total acidity, inosine monophosphate, alcohols, and aldehydes decreased. Correlation analysis and machine learning found that the level of histamine, total acidity and TVB-N could more accurately reflect shrimp sauce spoilage. Moreover, acetic acid and isoleucine were closely associated with flavor deterioration, while nonanal was strongly linked to odor decline. These findings clarified the relationship between physicochemical changes and flavor evolution, providing a foundation for establishing a robust and enhanced quality evaluation system for low-salt shrimp sauce.