Univariate Decision TreeIt refers to a decision with only one variable, that is, each time a node is split, only one feature in the feature set will be selected. This indicates that the classification boundary of the decision tree has multiple segments, which are composed of several segments parallel to the coordinate axis.

The univariate decision tree algorithm results in a large tree size, complex rules and difficulty in understanding, so a multivariate decision tree is generally used to solve the problem.

Unlike univariate decision trees, the learning of multivariate decision trees is not to find the optimal attribute for each non-node, but to try to build a suitable linear classifier.

Related terms: decision tree, multivariate decision tree

References

【1】<http://bealin.github.io/2017/02/27/Machine Learning Series>—5-Decision Tree/

【2】http://www.jsjkx.com/jsjkx/ch/reader/view_abstract.aspx?file_no=26283084&flag=1