Haiyan Zhao1, Feng Wang1, Qingli Yang1. 1. College of Food Science and Engineering, Qingdao Agricultural University, Qingdao, P. R. China.
Abstract
BACKGROUND: Multi-elements have been widely used to identify the geographical origins of various agricultural products. The objective of this study was to investigate the feasibility of identifying the geographical origins of peanut kernels at different regional scales by using the multi-element fingerprinting technique. The concentrations of 20 elements [boron (B), magnesium (Mg), phosphorus (P), potassium (K), calcium (Ca), etc.] were determined in 135 peanut samples from Jilin Province, Jiangsu Province, and Shandong Province of China. Data obtained were processed by one-way analysis of variance (ANOVA), principal components analysis (PCA), k nearest neighbors (k-NN), linear discriminant analysis (LDA), and support vector machine (SVM). RESULTS: Peanut kernels from different regions had their own element fingerprints. The k-NN, LDA, and SVM were all suitable to predict peanut kernels according to their grown provinces with the total correct classification rates of 91.2%, 91.1%, and 91.1%, respectively. While SVM was the best to identify different grown cities of peanut kernels with the prediction accuracy of 91.3%, compared to 72.2% and 78.3% for k-NN and LDA, respectively. CONCLUSION: It was an effective method to identify producing areas of peanut kernels at different regional scales using multi-element fingerprinting combined with SVM to enhance regional capabilities for quality assurance and control.
BACKGROUND: Multi-elements have been widely used to identify the geographical origins of various agricultural products. The objective of this study was to investigate the feasibility of identifying the geographical origins of peanut kernels at different regional scales by using the multi-element fingerprinting technique. The concentrations of 20 elements [boron (B), magnesium (Mg), phosphorus (P), potassium (K), calcium (Ca), etc.] were determined in 135 peanut samples from Jilin Province, Jiangsu Province, and Shandong Province of China. Data obtained were processed by one-way analysis of variance (ANOVA), principal components analysis (PCA), k nearest neighbors (k-NN), linear discriminant analysis (LDA), and support vector machine (SVM). RESULTS:Peanut kernels from different regions had their own element fingerprints. The k-NN, LDA, and SVM were all suitable to predict peanut kernels according to their grown provinces with the total correct classification rates of 91.2%, 91.1%, and 91.1%, respectively. While SVM was the best to identify different grown cities of peanut kernels with the prediction accuracy of 91.3%, compared to 72.2% and 78.3% for k-NN and LDA, respectively. CONCLUSION: It was an effective method to identify producing areas of peanut kernels at different regional scales using multi-element fingerprinting combined with SVM to enhance regional capabilities for quality assurance and control.