Regression Tree Splitting Criterion
Jump to navigation
Jump to search
A Regression Tree Splitting Criterion is a decision tree splitting criterion that can be used by a regression tree learning algorithm (to train a regression tree).
- Context:
- It can be constructed by using a Recursive Partitioning Algorithm.
- Example(s):.
- maximize the reduction in Sum of Squares.
- …
- Counter-Example(s):
- See: Regression Tree Pruning Algorithm.
References
1999
- (Torgo, 1999) ⇒ Louis Torgo. (1999). “Inductive Learning of Tree-based Regression Models." Ph.D. Thesis, Thesis, Faculty of Sciences, University of Porto
- We will now present an algorithm that finds the best split for continuous variables using the results of Definition 3.2. Assuming that we have a set of nt cases whose sum of the Y values is St, Algorithm 3.3 obtains the best split on a continuous predictor variable Xv.
- We will now present an algorithm that finds the best split for continuous variables using the results of Definition 3.2. Assuming that we have a set of nt cases whose sum of the Y values is St, Algorithm 3.3 obtains the best split on a continuous predictor variable Xv.
