Determination of Quality Parameters for Pacific Whiting Surimi Using Three Different Computer Analyses

Quality parameters for surimi from Pacific Whiting were detennined. Three different computer analyses were used: these included a traditional multiple linear regression, a model that documents the quality issues in a non-linear neural network framework, and an induction-type machine learning system to establish relationships of the on-board handling, physical characteristics and dockside storage condition on product quality. The strengths and weaknesses of each model are discussed.