Decision trees are used in a wide variety of areas such as radar signals, medical diagnostics, expert systems, remote sensing and voice recognition, to name just a few, successfully. Its ability to break a difficult decision-making procedure into a collection of simpler decisions and thus provide a solution is the most important feature of Decision Trees (Safavian & Landgrebe, 1991). The advantages of decision trees include: it enables to conduct feature selection implicitly, to discovering interactions and nonlinear relationships, it only requires little effort to prepare data without variable scaling, there is a reasonable number of missing values and it is not affected by outliers, it is easy to understand and to explain, it enables creating rules to formalize experts’ knowledge, a decision tree can be taken as a preprocessing step in the division of a dataset into small subsets, decision trees generate simple if-then statements, its subsets are purer than the original data set in relation to the target variable (Floares, Calin & Manolache, 2016; Kim, 2016).
We have drawn decision trees for the classification problem using German credit data and for a regression problem to predict housing SalePrice (this dataset is taken from Kaggle and attached to this discussion with the R script used to produce the plots).
Floares, A. G., Calin, G. A., & Manolache, F. B. (2016, June). Bigger Data Is Better for Molecular Diagnosis Tests Based on Decision Trees. In International Conference on Data Mining and Big Data (pp. 288-295). Springer, Cham.
Kim, K. (2016). A hybrid classification algorithm by subspace partitioning through a semi-supervised decision tree. Pattern Recognition, 60, 157-163.
Safavian, S. R., & Landgrebe, D. (1991). A survey of decision tree classifier methodology. IEEE transactions on systems, man, and cybernetics, 21(3), 660-674.