Use of machine learning to establish limits in the classification of hyperaccumulator plants growing on serpentine, gypsum and dolomite soils

The so-called hyperaccumulator plants are capable of storing hundred or thousand times bigger quantities heavy metals than normal plants, which makes hyperaccumulators very useful in fields such as phytoremediation and phytomining. Among these there many serpentinophytes, i.e., that grow exclusively on ultramafic rocks produce soils with a great proportion metals. Even though multiple classifications, the lack consensus regarding parameters to use determine whether plant is hyperaccumulator, well arbitrariness stablished thresholds, bring about need propose more objective criteria. To this end, mineral composition data from different vegetal species were analysed using machine learning techniques. Three complementary case studies established. Firstly, classified three types soils: dolomite, gypsum serpentine. Secondly, Ni find differentiated subgroups. Lastly, association carried out soil type. Results classification task reach success rate over 75%. Clustering by concentration parts per million (ppm) resulted four groups cut-off points 2.25, 100 (accumulators) 3000 ppm (hyperaccumulators). Associations confidence level above 90% found between high levels serpentine soils, Zn same type soil. Overall, work demonstrates potential analyse composition. Finally, after consulting red list IUCN those countries richness species, it evident greater effort should be made establish conservation status flora.