An explainable machine learning model for identifying geographical origins of sea cucumber Apostichopus japonicus based on multi-element profile

Abstract

The geographical origin of sea cucumber Apostichopus japonicas plays an important role in determining its market value. This study investigated the feasibility of using multi-element profile combined with explainable machine learning to trace the origin of sea cucumber in China. Multi-element profile (23 elements) of 167 sea cucumber samples was determined with ICP-OES and ICP-MS, and used for construction and evaluation of 4 ensemble learning models. Extreme gradient boosting (XGBoost) model achieved superior performance with an overall accuracy, precision, recall, F1 score and AUC as 0.95, 0.93, 0.91 and 1, respectively. The Shapley Additive Explanations (SHAP) algorithm was subsequently applied to interpret the XGBoost model output for desirable geographical information. Se was identified as the most important elemental marker for discriminating sea cucumber origins. Therefore, with clarified scientific support, multi-element profile combined with machine learning model could serve as a powerful tool for identifying the provenance of sea cucumber.