Association Rule Mining of Mushroom Toxicity Based on the FP-Growth Algorithm

Mushroom poisoning remains a significant public health concern in China, with frequent outbreaks in recent years. To systematically uncover complex toxicity patterns from large-scale mushroom datasets, this study utilizes the publicly available Secondary Mushroom Dataset and implements the Frequent Pattern Growth(FP-Growth) to perform association rule mining on mushroom toxicity. Comprehensive data preprocessing and feature engineering were conducted to transform morphological and ecological characteristics into a transactional format suitable for analysis. The results reveal that specific feature combinations—such as close gill spacing, absence of bruising, sticky cap surface, presence of a zone-type ring, yellow stipe, autumn collection, and meadow habitat—exhibit exceptionally high predictive value for toxicity risk, with some rules achieving a confidence of 1.0 and a lift greater than 6.5. These findings provide robust scientific evidence and practical guidelines for mushroom toxicity identification, offering valuable support for digital mushroom identification systems, food safety regulation, and public health emergency response. Furthermore, this work demonstrates the advantages of the FP-Growth algorithm in efficiently mining high-dimensional biological data, thus laying a methodological foundation for future studies in bioinformatics and food safety analytics.