Fig. 17
From: Constructing phylogenetic networks via cherry picking and machine learning
Classification tree with depth 4 of (a) the normal data set and (b) the LGT data set. For each node in the trees, except for the terminal ones, the first line is the feature condition. If this condition is met by a data point, it traverses to the left child node, otherwise to the right one. In the terminal nodes this line is omitted as there is no condition given. In each node, as also indicated with labels in the root node, the second line ‘samples’ is the proportional number of samples that follow the YES/NO conditions from the root to the parent of that node during the training process. The ‘value’ list gives the proportion of data points in each class, compared to the sample of that node. The last line indicates the most dominant class of that node. If a data point reaches a terminal node, the observation will be classified as the indicated class