FBP: A Frontier-Based Tree-Pruning Algorithm
FBP: A Frontier-Based Tree-Pruning Algorithm
复制标题
FBP：基于边界的树剪枝算法
DOI：
--
复制
发表时间：
2006
影响因子：
2.1
通讯作者：
Shuchun Wang
中科院分区：
文献类型：
--
作者：
X. Huo;Seoung Bum Kim;K. Tsui;Shuchun Wang
提出了一种基于边界的树修剪算法(FBP)。新方法具有与代价-复杂性剪枝(CCP)相当的计算复杂性。关于树修剪，它提供了全方位的信息：(1)给定惩罚参数λ的值，它给出由复杂性惩罚方法指定的决策树；(2)给定决策树的大小，它提供惩罚参数λ的范围，在该范围内，复杂性惩罚方法呈现该树大小；(3)它找到不可接受的树大小-无论惩罚参数的值是多少，基于复杂性惩罚框架得到的树将永远不会具有这些大小。对真实数据集的模拟显示了一个“惊喜”：在复杂性惩罚方法中，大多数树的大小是不可接受的。FBP有助于更忠实地实现交叉验证(CV)，这是模拟所青睐的。利用前馈神经网络，分析了循环伏安法的稳定性。
A frontier-based tree-pruning algorithm (FBP) is proposed. The new method has an order of computational complexity comparable to cost-complexity pruning (CCP). Regarding tree pruning, it provides a full spectrum of information: namely, (1) given the value of the penalization parameter λ, it gives the decision tree specified by the complexity-penalization approach; (2) given the size of a decision tree, it provides the range of the penalization parameter λ, within which the complexity-penalization approach renders this tree size; (3) it finds the tree sizes that are inadmissible---no matter what the value of the penalty parameter is, the resulting tree based on a complexity-penalization framework will never have these sizes. Simulations on real data sets reveal a “surprise:” in the complexity-penalization approach, most of the tree sizes are inadmissible. FBP facilitates a more faithful implementation of cross validation (CV), which is favored by simulations. Using FBP, a stability analysis of CV is proposed.