Machine Learning Models to Classify Shiitake Mushrooms (Lentinula edodes) According to Their Geographical Origin Labeling

The shiitake mushroom has become the second most consumed mushroom globally, offering numerous health benefits, but its origin is often unclear, which concerns consumers. This study developed machine learning algorithms to determine the geographical origin of shiitake mushrooms consumed in Korea using experimental data (δ13C, δ15N, δ18O, δ34S, and origin). When classifying shiitake into three categories (Korean, Chinese, and those from Chinese inoculated sawdust blocks), the random forest model showed the highest accuracy (0.940) and kappa value (0.908). For two categories (Korean and Chinese, with the latter including mushrooms from Chinese sawdust blocks), the support vector machine model excelled with an accuracy of 0.988 and kappa of 0.975. In a different two-category classification (Korean and Chinese, but with Chinese sawdust block mushrooms in the Korean category), the random forest was again the best model. Testing phase accuracy ranged from 0.839 to 0.964, indicating the potential for real-world application.

Authors: Raquel Rodríguez-Fernández, Ángela Fernández-Gómez, Juan C. Mejuto and Gonzalo Astray.

To find out more details, you can read the full paper in Foods.