An expert system for species discrimination and grade identification of fish maw

Abstract

Fish maw, derived from dried swim bladders of fish, is valued for its nutritional and medicinal properties, which has led to an increased market demand. However, price variability based on species and grades has results in unethical practices such as counterfeiting and mislabeling, highlighting the need for reliable quality authentication. To address this, an expert system using MATLAB software has been developed. This system employs nuclear magnetic resonance (NMR) technology and pattern recognition methods to identify fish maw species and classify their grade. The study analyzed ten species across three grades of fish maw, identifying 43 nutritional components from NMR spectra, including sugars, amino acids, fatty acids, organic acids, and vitamins. Univariate statistical analysis was integrated with multivariate statistical analyses, including principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), and orthogonal partial least squares (OPLS-DA) using SIMCA software, to develop models for species and grades identification. A four-dimensional volcano map was constructed to highlight characteristic components of various fish maws, resulting in an NMR database for common fish maws. The expert system’s accuracy was validated with new samples, achieving 92.6% for species identification and 90.0% for grade classification. This study provides a valuable tool for the quality evaluation of fish maw and a scientific basis for market regulation.