METODE KLASIFIKASI BERSTRUKTUR POHON DENGAN ALGORITMA CRUISE

Abstract

Two univariate split methods are proposed for the construction of classification trees with multyway splits named
CRUISE (Classification Rule with Unbiased Interaction Selection and Estimation). A major strenght of the univariate split
methods is that they have negligible bias in variable selection, both when the variables differ in the number of splits they offer
and when they differ in number of missing values. This is an advantage because inference from the tree sructures can be
adversely affected by selection bias. These methods also improve interpretability of trees by reducing tree depht.
Application of CRUISE algorithms to Fisher’s Iris data is to predict the variety of an Iris flower based on its petal and sepal
lenght and widht. Results show that it only takes one variable to do so. Therefoce, the new methods are highly competitive in
terms of computational speed and classification accuracy of future observation.