Combining Knowledge-and Data-Driven Fuzzy Approach to Evaluate Shelf-Life of Various Seafood Products

Due to the complexity of the deterioration process of seafood products,relying on one indicator is not adequate to determine the quality of such products.Usually,shelf-life was estimated based on various indicators complicating the decision-making process.Decision Support Systems are considered as a good solution.The current study aims to establish a simple and novel fuzzy model based on a combination of knowledge-and data-driven approaches to define a fuzzy quality deterioration index(FQDI)in various seafood products(rainbow trout,threadfin bream,and white shrimp samples)during cold storage.Total volatile basic nitrogen(TVB-N)and psychrotrophic microorganisms counts(PMCs)were determined based on traditional methods.The sensory analysis was performed by a data-driven fuzzy approach.Overall,the shelf-life of the rainbow trout fillet was estimated to be 8 d,based on all the freshness parameters.However,the shelf-life of the Japanese threadfin bream fillet was 5-7 d according to the microbial and chemical parameters,respectively.This time for shrimp samples was 6-8 d using sensory score and TVB-N contents.The results of data-driven fuzzy approach showed all of the quality properties were considered as the'Important'-'Very Important'(defuzzification score>75).The TVB-N and PMCs were the most and weakest freshness quality properties(defuzzified-values:84.64 and 78.75,respectively).Based on FQDI,the shelf-life of the rainbow trout,Japanese threadfin bream,and shrimp samples were estimated to be 8,5,and 7 d,respectively.This method was able to successfully provide a comprehensive deterioration index for evaluating the seafood shelf-life.Such a total index can be considered as a comprehensive output(y variable)to predict seafood freshness by rapid and nondestructive method.