important feature of CART is its ability to generate regression trees. Regression trees are trees where their leaves predict a real number and not a class. In case of regression, CART looks for splits that minimize the prediction squared error (the least–squared deviation). The prediction in each leaf is based on the weighted mean for node. CHAID (CHi-squared Automatic Interaction Detector) Performs multi-level splits when computing classification
From wind-swept seeds to efficient flight mechanism, all animals and plants have evolved a variety of dispersal abilities and have an important role in understanding species survival and evolution. Dispersal is largely driven by needs for resource acquisition particularly to seek suitable breeding habitats. Individuals dispersing between populations aid in gene spread and as such dispersal is subject to natural selection. Three main evolutionary drivers are thought to result in dispersal - habitat