TraverseTreeToMeasureClassificationQuality: FUNCTION: TEST PERFORMANCE OF CLUSTER TREE ON LABELED DATA...

In thecomeonman/CURD: Implementation of the CURD algorithm for clustering

Description

FUNCTION: TEST PERFORMANCE OF CLUSTER TREE ON LABELED DATA This function will traverse the tree to various depths and create a dataset with a row for each node in the tree. This dataset will contain details of how many points that node contains, whether or not it is a leaf node and what the proportion of various classes is on that node. Further, on the assumption that a node is assigned to a particular class, it tells us how accurate each node is (based on the impurity level) Depth-first traversal of the tree happens

Usage

TraverseTreeToMeasureClassificationQuality(dtTreeSummary, dtDataset, lTree,
  cClassLabelFeat)

Arguments

dtTreeSummary	todo
dtDataset	todo
lTree	todo
cClassLabelFeat	todo

Value

todo

thecomeonman/CURD documentation built on May 20, 2019, 7:37 a.m.