Figure 2.

The Predictive Clustering Tree (PCT) for multi-label classification (MLC), predicting the presence of fungal functional groups, isolated from beach sand and seawater. The decision process starts by separating samples based on their type (sand or seawater). From there on, Monthly rainfall rate, Monthly sunshine hours, Season, pH, Average air humidity and Average daily sunshine minutes are used to determine the prediction of the tree. For example, if samples come from seawater and their pH is higher than 8.11, the predicted labels (functional groups) are: Core genus, Human colonisers, Psychrotolerant fungi, and Halotolerant/halophilic fungi.