A novel composite colorimetric sensor array for quality characterization of shrimp paste based on indicator displacement assay and etching of silver nanoprisms

Abstract In this study, a novel amino acids and salts‐sensitive colorimetric sensor array (CSA) was constructed to evaluate shrimp paste quality based on mechanisms of indicator displacement assay (IDA) and silver nanoprisms (AgNPRs) etching. Three supervised learning methods including linear discriminant analysis (LDA), K‐nearest neighbor (KNN), and Support Vector Machine (SVM) were applied to qualitatively distinguish shrimp paste from different geographical origins with the discriminant accuracy for prediction set of 94.44%, 97.22%, and 100%, respectively. Partial least squares (PLS) and SVM were further used to quantitatively predict the crucial compounds in shrimp paste. The correlation coefficients for prediction set ( R p ) of amino acid nitrogen and salt using PLS model were 0.8875 and 0.9478, respectively. And the prediction performance was significantly improved by using SVM analysis with R p of 0.9312 and 0.9500, respectively. The results indicated that the CSA can be an effective tool in the quality characterization of shrimp paste. Practical applications A novel amino acids and salts‐sensitive colorimetric sensor array (CSA) was constructed to evaluate shrimp paste quality based on mechanisms of indicator displacement assay (IDA) and silver nanoprisms (AgNPRs) etching. This study combined CSA with pattern recognition methods including LDA, KNN, and SVM modeling methods to effectively identify six different shrimp pastes with the highest recognition rate of the SVM model (100% for both training set and prediction set). Partial least squares (PLS) and SVM modeling methods were further applied to quantitatively predict the amino acid nitrogen and salt content of six different types of shrimp pastes with an excellent prediction performance.