Premium
Novel prediction of heavy metal residues in fish using a low‐cost optical electronic tongue system based on colorimetric sensors array
Author(s) -
Han Fangkai,
Huang Xingyi,
Teye Ernest
Publication year - 2019
Publication title -
journal of food process engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.507
H-Index - 45
eISSN - 1745-4530
pISSN - 0145-8876
DOI - 10.1111/jfpe.12983
Subject(s) - partial least squares regression , electronic tongue , extreme learning machine , fish <actinopterygii> , chemistry , correlation coefficient , metal , chemometrics , biological system , environmental science , computer science , chromatography , mathematics , artificial neural network , artificial intelligence , fishery , statistics , food science , organic chemistry , taste , biology
For simultaneous and quantitative prediction of three toxic heavy metal residues (Pb, Cd and Hg) in fish, the novel and low‐cost colorimetric sensors array was fabricated. Forty eight fish samples (Carassius carassius) retailed in Zhenjiang, China, were used. Inductively coupled plasma mass spectrometry was employed as the reference analytical method. The partial least square regression (PLS) and extreme learning machine (ELM) respectively were comparatively used for data modeling. Results showed that the ELM models built were better than the PLS models with correlation coefficient in the prediction set for Pb, Cd, and Hg as 0.854, 0.83, and 0.845 respectively; while, the corresponding root mean squares error were 0.102 mg/kg, 0.026 mg/kg, and 0.016 mg/kg respectively. This work demonstrates that the developed artifical gustatory system has a great potential in the simultaneous and quantitative prediction of heavy metal residues in fish and could be used perhaps for other fish products. Practical applications Fish is an important food commodity worldwide. It is very vital to monitor heavy metals residues in raw fishes to guarantee food safety. Traditional analytical methods for heavy metals determination in fish are quite cumbersome and usually expensive. The novel low‐cost and simple colorimetric sensors array based on pyridylazo and porphyrin indicators has been developed in the present study. Results showed that, photoelectric electronic tongue based on colorimetric sensors coupled with extreme learning machine model has a great potential in the simultaneous and quantitative prediction of toxic heavy metal residues in fish.