Assessment of Frozen Stored Silver Carp Surimi Gel Quality Using Synthetic Data-Driven Machine Learning (SDDML) Model

Area of Science:

• Food Science and Technology
• Biotechnology
• Machine Learning Applications

Background:

• Invasive Silver Carp (Hypophthalmichthys molitrix) is a potential surimi resource in North America.
• Frozen storage significantly degrades Silver Carp surimi's gel-forming ability, impacting its value-added potential.

Purpose of the Study:

• To investigate the effects of frozen storage duration on Silver Carp surimi gel properties.
• To evaluate the efficacy of food additives in mitigating frozen storage-induced deterioration.
• To develop and apply a synthetic data-driven machine learning (SDDML) approach for optimizing surimi gel quality.

Main Methods:

• Analysis of Silver Carp surimi gel properties after six months of frozen storage.
• Testing the effectiveness of food additives including microfiber, transglutaminase, and chicken skin collagen.
• Development and validation of a random forest-based SDDML model to predict gel quality attributes.

Main Results:

• Short-term frozen storage (<2 months) improved surimi gelability; extended storage (>2 months) was detrimental.
• Transglutaminase (0.1 wt%) was the most effective additive, restoring gel quality comparable to fresh samples.
• The SDDML model accurately represented experimental data (R² = 0.871-0.889) and improved prediction of gel quality attributes.

Conclusions:

• Transglutaminase can effectively counteract the negative effects of long-term frozen storage on Silver Carp surimi.
• SDDML offers a promising approach to overcome data scarcity in optimizing food product formulations.
• The study demonstrates a pathway to enhance the utilization of invasive species like Silver Carp into high-value food products.